Multi-task modular hierarchical regulation and control system based on multi-core CPU architecture

Abstract

The invention discloses a multi-task modular hierarchical regulation and control system based on a multi-core CPU architecture, which relates to the technical field of power generation control, and comprises a domestic multi-core CPU hardware basic module adopting a Feiteng or Peng architecture, a triple modular redundancy design and a 24-bit AD signal acquisition module based on a domestic operating system; the multi-task hierarchical scheduling core module is used for five-level priority division, a core binding and dynamic migration strategy and an improved round-robin algorithm; the modular function adaptation interface module supports hot plug, parameter self-matching and dual-path redundancy switching; the running state monitoring and feedback module is used for evaluating the health state in a weighted fusion manner; and an open development support module, dual-mode configuration and an encapsulation algorithm support third-party development and version management. According to the method, DCS autonomous controllability and efficient and balanced multi-task scheduling are realized, the operation monitoring and operation and maintenance intelligent level is improved, and safe and efficient operation of a unit is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power generation control, and in particular to a multi-task modular hierarchical regulation system based on a multi-core CPU architecture. BACKGROUND

[0002] With the transformation of the thermal power industry towards clean and efficient, flexible and intelligent, the distributed control system (DCS) as the core hub of the unit operation, the demand for autonomous controllability and multi-task processing ability is increasingly urgent. The traditional DCS system relies on imported CPU and operating system, the core hardware and software architecture is limited by external technology, and the upgrade cost is high, the expansibility is insufficient, which is difficult to adapt to the demand of domestic unit for customized service and large-scale monitoring expansion. Although the early domestic controller is gradually applied, there are still problems such as low single-core computing power, slow interrupt response speed and delayed complex calculation and processing, which cannot meet the high-frequency and high-concurrency task processing demand of large-capacity units above 300 MW.

[0003] The limitations of multi-task scheduling and module adaptation further restrict the performance improvement of DCS system. The traditional system adopts a single-level task scheduling mode, and the task priority is not accurately divided. The core control task of the unit (such as boiler water level regulation) and the idle data statistical task compete for CPU resources, which easily leads to delay of key control task and affects the stability of unit operation. At the same time, the traditional DCS system is highly closed, the interface type is single and lacks standardized design, and third-party developers are difficult to access for model and algorithm development. Intelligent power generation new technologies such as predictive control and neural network algorithm cannot be effectively integrated, forming an information island, leading to system function iteration lagging behind industry demand. When the module is accessed, the parameters need to be manually debugged, which is low in adaptation efficiency and easy to cause compatibility conflict, increasing the operation and maintenance cost.

[0004] The intelligent level of operation monitoring and operation and maintenance management is also difficult to meet the modern power generation demand. The state monitoring of the traditional DCS system is mostly limited to basic parameter acquisition, lacking fusion analysis of multi-dimensional data such as CPU load, task response time and module temperature, and unable to quantitatively evaluate the system health status. The fault early warning relies on manual judgment, which is easy to miss or misjudge in the face of massive operation parameters, and it is difficult to trace back after the fault occurs, which requires the experience of operation and maintenance personnel to troubleshoot, and the processing cycle is long. In addition, the system security protection mostly stays at the basic account password level, lacking full-level protection of hardware data, transmission process and access authority, being easy to be affected by unauthorized operation and data leakage risk, and unable to guarantee the safety of core control data. These problems together cause the traditional DCS system to be unable to fully adapt to the comprehensive demand of modern power industry for autonomous controllability, efficient scheduling, open compatibility and intelligent operation and maintenance. SUMMARY

[0005] The multi-task modular hierarchical regulation system based on the multi-core CPU architecture is provided to solve the problems in the prior art.

[0006] To achieve the above object, the application adopts the following technical scheme: a multi-task modular hierarchical regulation system based on a multi-core CPU architecture, comprising: The domestic multi-core CPU hardware basic module is optional in Fengteng four-core or Kunpeng eight-core architecture, supports pin and instruction set compatible upgrade, integrates three-mode redundancy design, the main and standby switching is synchronous through optical fiber, the arbitration unit checks data consistency, is equipped with a collection module, double photoelectric isolation + metal shielding, anti-interference reaches IEC standard, supports signal access, based on Kirin V10 or UOS, supports monitoring expansion; The multi-task hierarchical scheduling core module constructs a five-level scheduling system: emergency interrupt level, real-time control level, communication interaction level, system supervision level, and idle optimization level; adopts the "core binding + dynamic load migration" strategy, binds the emergency / real-time task to the dedicated core, and the rest of the tasks are rotated according to the improved time slice; The modular function adaptive interface module is preset with five kinds of general interfaces, supports four kinds of data format analysis; adapts to 12 kinds of function modules, the interface is standardized, supports hot plug, and automatically matches parameters; double-channel redundancy design, LED indicator light feedback state; The running state monitoring and feedback module collects key parameters in real time, generates a health score by a weighted fusion algorithm, displays data and curves on a touch screen, and supports multi-dimensional filtering; The open development support module is a dual-mode configuration platform, provides a drag-and-drop component, encapsulates an intelligent algorithm classification warehouse, supports model simulation and release, can version backtracking and log recording, and third parties can access through SDK and upload custom algorithms.

[0007] Further, it further comprises a task priority dynamic evaluation unit, and the evaluation method is ; wherein P is the task priority score; is the task timeliness weight; is the resource consumption weight; is the associated influence weight; T is the remaining time of the task deadline; C is the CPU resource occupation ratio required by the task; S is the importance score of the task associated equipment; the task execution order can be dynamically adjusted through the evaluation.

[0008] Further, it also includes a module running temperature self-adaptive adjustment unit; an integrated distributed temperature sensor array that collects temperature data in real time; a three-level regulation strategy of "active heat dissipation + passive heat dissipation + task regulation": when the temperature is greater than or equal to 65°C, start the intelligent cooling fan and adjust the rotating speed according to the temperature gradient; when the temperature is greater than or equal to 75°C, trigger the phase change heat dissipation layer of the passive heat dissipation fin to work; when the temperature is greater than or equal to 85°C, temporarily suspend the idle optimization level task, and distribute the real-time control level task to the low-temperature core.

[0009] Further, it also includes a third-party module compatibility verification unit, and the verification method is C=αN / βM; wherein C is the compatibility index; α is the success rate of third-party module adaptation; N is the total number of accessed third-party modules; β is the system inherent module adaptation coefficient; M is the total number of system inherent modules; a full-process automatic test mechanism is matched, and after the module is accessed, "basic function test, stress test, fault injection test" are automatically executed, the module response time, resource occupation rate and conflict frequency are recorded, and a compatibility report containing test data and optimization suggestions is generated.

[0010] Further, it also includes an intelligent energy consumption optimization unit, which is in data intercommunication with the CPU hardware basic module and the multi-task hierarchical scheduling core module; a "dynamic voltage frequency regulation + core hibernation + peripheral control" collaborative strategy is adopted: when the system load is less than or equal to 25%, automatically turn off half of the idle CPU cores; when the load is 25%-75%, maintain the core voltage at 1.2V; when the load is greater than or equal to 75%, start all cores; the unit generates an energy consumption analysis report every 5 minutes, compares the unit task energy consumption under different loads, and optimizes the parameter adjustment threshold value in combination with historical data.

[0011] Further, it also includes a CPU core load balancing unit, and the regulation method is ; wherein is the load balancing degree; is the load correction coefficient; is the load rate of the i-th CPU core; is the average load rate of all CPU cores; is the total number of CPU cores; through the calculation, the overloaded core is identified in real time, and the task migration process is automatically started: the idle optimization level task is preferentially migrated, followed by the communication interaction level task, the task state is saved through data snapshot technology before migration, and the task running state is checked after migration; for the real-time control level task, the "pre-migration + seamless switching" mode is adopted.

[0012] Further, it further includes a data breakpoint continuation unit; an integrated 16GB local cache area, which is divided into cache priorities according to'real-time control data > communication interaction data > system supervision data > idle time optimization data'; when the communication link is interrupted, the untransmitted data is automatically stored in the cache area according to the priority; after the link is restored, the priority continuation mechanism is started, and the real-time control data is preferentially continued, and the rest of the data is continued in time sequence; the unit is provided with a double data verification mechanism: CRC32 check code is generated before transmission, and the check code is compared after transmission; after the continuation is completed, an integrity report is generated, and the missing data and the supplement transmission situation are marked; the support and the cloud storage are linked, and when the local cache is about to overflow, the low-priority data is automatically uploaded to the cloud.

[0013] Further, it further includes a customized task configuration unit; a 10.1-inch touch visual configuration interface is provided, and three typical scene templates of 'thermal power unit control, heat network scheduling, equipment operation and maintenance' are preset; the configuration parameters cover task priority, execution period, CPU core binding, resource allocation ratio; the configuration parameters are stored in a non-volatile flash memory, and are automatically loaded after system restart; an AI parameter recommendation engine is built-in, three sets of parameter schemes are generated for selection based on the scene type, task type and device quantity input by the user, combined with historical optimal configuration data; at the same time, the parameter legality verification function is provided, when the configuration parameters exceed the hardware bearing range, the critical value is automatically prompted and recommended.

[0014] Further, it further includes a full-level security protection unit, which constructs a five-dimensional protection system of 'hardware-software-data-access-log'; the hardware layer adopts the national SM4 encryption chip; the software layer integrates an intrusion detection model based on deep learning; the data layer adopts the AES-256 encryption algorithm to guarantee the transmission safety, and the storage data adopts the sharding encryption + distributed storage; the access layer implements five-level permission management, and adopts 'account + password + fingerprint / facial biometric + USBKey' four-factor authentication; the log layer adopts the block chain storage technology, and records all operations and abnormal events.

[0015] Further, it further includes a remote diagnosis and operation and maintenance unit; equipped with 4G / 5G / Wi-Fi three-mode communication module, real-time uploading system running data to the cloud operation and maintenance platform; the operation and maintenance personnel can remotely view real-time data, historical trends, fault records, and issue parameter adjustment instructions by logging in the platform through the PC end or mobile end; when the system triggers a serious fault, the fault snapshot is automatically pushed to the operation and maintenance terminal, the remote start of the fault simulation is supported, and the operation and maintenance personnel can optimize and repair the scheme through simulation debugging.

[0016] Compared with the prior art, the present application has the following advantages: In terms of autonomous controllability and operation stability, the application is based on multi-core CPU architecture to build hardware foundation, supports mainstream domestic chip configuration such as Feiteng and Kunpeng, and is matched with domestic operating systems such as Kirin and United Credit, which completely gets rid of the dependence on imported technology. The integration of triple modular redundancy design and double anti-interference signal acquisition module greatly improves the system operation continuity and data acquisition accuracy, avoiding operation abnormalities caused by hardware failure and signal interference. The construction of five-level task scheduling system realizes the accurate hierarchical processing of tasks, and the core control task is bound to the exclusive CPU core, effectively solving the problem of resource contention and significantly improving the real-time response speed of the system, providing a solid foundation for the safe and stable operation of the unit.

[0017] The multi-task scheduling and load balancing capability is essentially improved, and the system processing efficiency is greatly optimized. Through dynamic evaluation of task priority and core load balancing control, the task execution order can be adjusted in real time according to the timeliness of the task, resource consumption and importance of the associated equipment, and overloaded core tasks can be automatically migrated to avoid single-core resource overload. The application of improved time slice round-robin algorithm reduces the task switching overhead, significantly enhances the single-core concurrent processing capability, and not only guarantees the priority execution of emergency interruption and real-time control tasks, but also efficiently processes communication interaction and idle optimization tasks, realizing balanced and efficient operation of various tasks.

[0018] The open compatibility feature breaks the closed barriers of traditional systems and significantly improves the system expandability. Standardized multi-type interfaces and hot plug design enable automatic parameter matching when modules are connected, eliminating the need for manual debugging and greatly improving adaptation efficiency and compatibility. The establishment of the "zero-code graphical + code-level secondary development" dual-mode configuration platform and the classification algorithm warehouse reduces the third-party development threshold, supports rapid development and deployment of customized modules, promotes the rapid landing of new intelligent power generation technologies, effectively avoids the information island problem, and provides the possibility for continuous iteration of system functions.

[0019] The operation monitoring and operation intelligentization level are significantly improved, and the operation cost is greatly reduced. Multi-dimensional parameter acquisition and weighted fusion analysis realize the quantitative evaluation of the system health status, the fault tracing function can quickly locate the abnormal source and generate processing suggestions, reducing the dependence on the experience of operation personnel. Remote diagnosis and operation unit support realize real-time data upload and remote debugging through industrial internet, and can quickly push snapshots and repair schemes when faults occur, combined with operation knowledge base to shorten the processing cycle. The full-level security protection system constructs barriers from five dimensions of hardware, software, data, access, and log, effectively preventing unauthorized operations and data leakage risks, and ensuring the safety of core control data.

[0020] Furthermore, the intelligent energy consumption optimization unit achieves a dynamic balance between energy consumption and performance by dynamically adjusting hardware parameters and core operating status, adapting to different loads and environmental scenarios, and reducing system operating costs. The customized task configuration unit presets typical scenario templates and provides AI parameter recommendations, lowering the user's operating threshold and flexibly adapting to the application needs of different industries such as power generation and metallurgy, further expanding the system's applicability. Attached Figure Description

[0021] Figure 1 This is a schematic block diagram of the multi-task modular hierarchical control system based on a multi-core CPU architecture proposed in this invention. Figure 2 This is a comparison chart of the task scheduling performance of traditional methods and the present invention; Figure 3 A comparison chart of system load balancing under different load conditions; Figure 4 This is a diagram showing the results of a third-party module compatibility assessment. Figure 5 This is a diagram illustrating the effect of the module's adaptive temperature adjustment during operation. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] 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.

[0024] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0025] Reference Figures 1 to 5 A multi-task modular hierarchical control system based on a multi-core CPU architecture, comprising: The domestically produced multi-core CPU hardware module offers optional configurations using either the Phytium FT-2000 / 4 quad-core or the Kunpeng 920 octa-core architecture, with power consumption of 45W and 65W respectively. It boasts a continuous trouble-free operating time of ≥18,000 hours and supports dual low-cost upgrades based on pin and instruction set compatibility. It integrates a triple-redundancy design (master / backup + arbitration unit), with master / backup switching synchronized via fiber optic cable, achieving a switching time ≤25ms. The arbitration unit verifies data consistency in real-time to prevent switching anomalies. Equipped with a high-performance, high-speed signal acquisition module, it employs a 24-bit AD converter with an adjustable sampling frequency of 100kHz-1MHz. It integrates dual opto-isolation (5000V isolation voltage) and a metal shielded cavity, achieving anti-interference capabilities up to the IEC61000-4-4 standard. A single module supports 64 channels of multi-type signal input, covering 4-20mA analog signals, DI / DO digital signals, 100Hz-20kHz pulse signals, and RS485 bus signals. A built-in automatic calibration circuit calibrates every 12 hours, with a sampling error ≤±0.05%. The underlying environment is built on the domestic Kylin V10 or UnionTech UOS operating system, supporting elastic expansion of 1,000-10,000 monitoring points, with a system startup time of ≤60 seconds.

[0026] The core module of multi-task hierarchical scheduling constructs a five-level task scheduling system, divided by priority into emergency interruption level, real-time control level, communication interaction level, system supervision level, and idle-time optimization level. The emergency interruption level responds to hardware faults with a response time ≤10ms; the real-time control level handles core unit control (such as boiler water level regulation) with a response time ≤20ms; the communication interaction level is responsible for cross-module data transmission with a latency ≤50ms; the system supervision level performs module status checks with a cycle ≤50ms; and the idle-time optimization level handles data statistics and model training, consuming ≤15% of resources. A "core binding + dynamic load migration" strategy is adopted, binding emergency interruption level and real-time control level tasks to dedicated CPU cores, while other tasks are scheduled using an improved time-slice round-robin algorithm. The single-core task concurrent processing capacity is ≥800 tasks / second, and the task switching overhead is ≤1ms.

[0027] The modular functional adapter interface module is pre-installed with five common interfaces: Ethernet, MODBUSRTU / TCP, OPCUA, MQTT, and DDS, with a data exchange rate of ≥200MB / s. It supports SQL, JSON, XML, and Protobuf data format parsing. It adapts to 12 types of functional modules, including intelligent monitoring, centralized monitoring, unit-level one-click start / stop, and energy consumption optimization. The interfaces adopt standardized mechanical structures and electrical protocols, support hot-swapping, and automatically trigger a parameter self-matching process (including signal type, data range, and communication protocol verification) upon module connection, with a matching time ≤2s. It features a built-in dual-redundant interface design; in case of a primary interface failure, redundant interface switching is triggered, with a switching time ≤40ms. Interface status is provided in real-time via LED indicators (green for operation, red for fault, and yellow for pending matching).

[0028] The operational status monitoring and feedback module collects 30 key parameters in real time, including CPU core load, task response time, module operating temperature, and interface communication rate, with a sampling interval of ≤50ms. A weighted fusion algorithm (hardware parameters weighted 0.6, software parameters weighted 0.4) generates a system health status score from 0 to 100. The system's operational data and trend curves for each module are displayed on a 7-inch touchscreen, supporting multi-dimensional filtering (time, module, parameter type). When a parameter deviates from a threshold of ±8%, an audible and visual alarm is triggered, with an alarm response time of ≤800ms. Simultaneously, the fault tracing process is automatically initiated, using cause-effect graph analysis to associate abnormal parameters with related tasks, generating a fault handling suggestion list including processing steps, required tools, and estimated time.

[0029] The open-source development support module adopts a dual-mode configuration platform of "zero-code graphical interface + code-level secondary development," providing a drag-and-drop component library (containing over 200 components for data acquisition, logic judgment, algorithm calling, and interface display). It encapsulates over 150 intelligent algorithms into a categorized repository, covering predictive control, fuzzy algorithms, neural networks, genetic algorithms, etc., categorized and labeled as "control, diagnostic, and optimization," with accompanying documentation explaining algorithm principles, applicable scenarios, and parameter ranges. It features built-in offline model simulation and online deployment capabilities. The simulation environment supports fault scenario simulation (such as CPU overload and interface interruption), automatically generating a version number after deployment (rule: year + month + serial number), supporting version rollback, difference comparison, and operation log recording (including developer, operation time, and modification content). Third-party developers can access the platform through a standardized SDK, customizing algorithm modules and uploading them to the algorithm repository.

[0030] This invention also includes a task priority dynamic evaluation unit for real-time optimization of task scheduling strategies, the evaluation method being... Where P is the task priority score, ranging from 0 to 100. The timeliness weight is set at 0.6 for emergency scenarios and 0.5 for routine scenarios; This is the resource consumption weight, with a value of 0.3. The weighting factor is 0.2; T represents the remaining time before the task deadline, in milliseconds; C represents the percentage of CPU resources required by the task, ranging from 0 to 1; and S represents the importance score of the associated equipment, ranging from 0 to 100 points (100 points for core equipment and 50 points for auxiliary equipment). This assessment allows for dynamic adjustment of the task execution order. For example, during unit start-up and shutdown, the priority of real-time control tasks can be increased to above 90 points to ensure that time-sensitive tasks and those associated with core equipment are executed first, avoiding delays in critical control tasks.

[0031] This invention also includes a module operating temperature adaptive adjustment unit to ensure hardware operational stability. An integrated distributed temperature sensor array (one per CPU core and one per functional module) collects temperature data in real time with a sampling interval of ≤100ms. A three-level adjustment strategy of "active cooling + passive cooling + task control" is adopted: when the temperature ≥65℃, the intelligent cooling fan is activated and its speed is adjusted according to the temperature gradient (50%-80% speed at 65℃-75℃); when the temperature ≥75℃, the phase change heat dissipation layer of the passive heat sink is triggered (phase change temperature 75℃); when the temperature ≥85℃, idle optimization tasks are temporarily suspended, real-time control tasks are allocated to the low-temperature core, and the CPU operating frequency is reduced to 80% of the baseline value. This unit automatically initiates a temperature calibration process every 24 hours, verifying sensor accuracy through a standard temperature source to ensure that the temperature monitoring error is ≤±0.8℃, avoiding adjustment anomalies caused by temperature misjudgment.

[0032] This invention also includes a third-party module compatibility verification unit to prevent system conflicts caused by module access. The verification method is C=αN / βM. Here, C is the compatibility index, with a value ≥0.93 indicating successful compatibility; α is the third-party module adaptation success rate, i.e., the ratio of the number of successfully running third-party modules to the total number of accessed modules, ranging from 0 to 1; N is the total number of accessed third-party modules; β is the system's inherent module adaptation coefficient, with a value of 0.97; and M is the total number of system's inherent modules. A fully automated testing mechanism is included. After module access, it automatically executes "basic function tests (1000 read / write operations), stress tests (simulating 10 times load), and fault injection tests (50 interface interruptions / data errors)," recording module response time (≤100ms), resource utilization (≤20%), and number of conflicts (≤1), generating a compatibility report containing test data and optimization suggestions to help third-party developers adjust module parameters (such as reducing data interaction frequency and optimizing memory usage).

[0033] This invention also includes an intelligent energy consumption optimization unit that communicates with the CPU hardware basic module and the multi-task hierarchical scheduling core module to dynamically adjust hardware parameters based on task load and operating scenario. It employs a collaborative strategy of "dynamic voltage and frequency adjustment + core hibernation + peripheral control": when the system load is ≤25%, it automatically shuts down half of the idle CPU cores, reduces the operating core voltage to 1.0V and the frequency to 60% of the baseline value, and simultaneously shuts down the power supply to interfaces not connected to modules; when the load is 25%-75%, it maintains the core voltage at 1.2V and the frequency at the baseline value, waking up cores according to the load ratio; when the load is ≥75%, it starts all cores, increasing the voltage to 1.3V and the frequency to 110% of the baseline value to ensure task processing efficiency. This unit generates an energy consumption analysis report every 5 minutes, comparing the unit task energy consumption under different loads and optimizing parameters and adjusting thresholds based on historical data. For example, in winter, the hibernation start-up load threshold is reduced to 20% (utilizing environmental heat dissipation to reduce energy consumption), while in summer it is maintained at 25% (avoiding frequent start-stop cycles that affect stability), achieving a dynamic balance between energy consumption and performance.

[0034] This invention also includes a CPU core load balancing unit to avoid single-core resource overload. The control method is L=γ∑(Ci-Cavg)² / n. Where L is the load balancing degree, with a value ≤0.08 indicating balance; γ is the load correction coefficient, with a value of 1.1; Ci is the load rate of the i-th CPU core, with a value of 0-1; Cavg is the average load rate of all CPU cores, with a value of 0-1; and n is the total number of CPU cores. This calculation identifies overloaded cores in real time (Ci≥Cavg+0.2) and automatically initiates the task migration process: prioritizing the migration of idle-time optimization tasks, followed by communication interaction tasks. Before migration, task status is saved using data snapshot technology; the interruption time during migration is ≤8ms; and after migration, the task running status (such as data integrity and response time) is verified to ensure no migration anomalies. For real-time control-level tasks, a "pre-migration + seamless switching" mode is adopted, loading the task environment in advance on the target core and completing the switch during the original core task execution interval to avoid control interruption.

[0035] This invention also includes a data interruption resumption unit to ensure the integrity of task data transmission, especially suitable for communication fluctuation scenarios in industrial environments. It integrates a 16GB local high-speed cache (non-volatile), prioritizing data according to the order of "real-time control data > communication interaction data > system monitoring data > idle-time optimization data". When the communication link is interrupted, untransmitted data is automatically stored in the cache according to priority, while marking the data generation time and the task to which it belongs. After the link is restored, a priority resumption mechanism is activated, prioritizing the real-time control data (transmission rate ≥ 50MB / s), with the remaining data resumed in chronological order. This unit has a built-in dual data verification mechanism: a CRC32 checksum is generated before transmission, and the checksum is compared after transmission; if verification fails, transmission is re-established. After resumption, an integrity report is generated, marking missing data and re-transmission status. It supports linkage with cloud storage; when the local cache is about to overflow, low-priority data is automatically uploaded to the cloud, freeing up local space (cloud transmission uses HTTPS encryption).

[0036] This invention also includes a customized task configuration unit, allowing users to adjust task parameters according to different application scenarios such as power generation and metallurgy, thus lowering the professional threshold. It provides a 10.1-inch touchscreen visual configuration interface with three preset typical scenario templates: "thermal power unit control," "heat network scheduling," and "equipment operation and maintenance," which users can directly select or customize. Configurable parameters include task priority (0-100 points), execution cycle (adjustable from 1ms to 1h), CPU core binding (optional "automatic allocation" or "manual binding"), and resource allocation ratio (adjustable CPU utilization from 0-50%). Configuration parameters are stored in non-volatile flash memory (1GB capacity) and automatically loaded after system restart. A built-in AI parameter recommendation engine generates three parameter schemes for selection based on user-input scenario type, task type, and number of devices, combined with historical best configuration data. It also features parameter validity verification; when configuration parameters exceed the hardware's carrying capacity (e.g., CPU utilization > 50%), it automatically prompts and recommends a critical value (e.g., 45%) to prevent system lag and task loss due to incorrect configuration.

[0037] This invention also includes a multi-layered security protection unit, constructing a five-dimensional protection system encompassing hardware, software, data, access, and logs to meet the security requirements of industrial control systems. At the hardware level, a national standard SM4 encryption chip (integrated into the CPU motherboard) is used to encrypt and store CPU operating data and keys. The chip is physically tamper-proof (data is destroyed upon disassembly). At the software level, a deep learning-based intrusion detection model is integrated, trained with over 100,000 industrial attack samples, achieving an accuracy of ≥99.9% in identifying abnormal access and malicious code injection, automatically isolating suspicious processes upon triggering. At the data level, AES-256 encryption ensures secure transmission, and data storage employs fragmented encryption and distributed storage (each fragment uses a different encryption key). At the access level, five-level access control is implemented (system administrator, operations engineer, developer, operator, and visitor), using four-factor authentication: account + password + fingerprint / facial biometrics + USBKey. After five failed authentication attempts, the account is automatically locked and an alarm is pushed. At the log level, blockchain storage technology is used to record all operations and abnormal events (unalterable), supporting a traceability period of ≥3 years.

[0038] This invention also includes a remote diagnostic and maintenance unit, supporting remote management and control throughout the entire lifecycle via the Industrial Internet, reducing on-site maintenance costs. Equipped with a 4G / 5G / Wi-Fi tri-mode communication module, it uploads system operation data (health score, task scheduling status, module parameters) to the cloud-based maintenance platform in real time, with an adjustable upload frequency of 1-10 seconds. Maintenance personnel can log in to the platform via PC or mobile terminal to remotely view real-time data, historical trends, and fault records, and issue parameter adjustment commands (such as modifying task priorities and calibrating signal acquisition modules). Command transmission uses end-to-end encryption. When the system triggers a serious fault (health score < 60 points), a fault snapshot (including abnormal parameters, task status, and environmental data) is automatically pushed to the maintenance terminal, supporting remote initiation of fault simulation (reproducing fault scenarios based on a digital twin model). Maintenance personnel can optimize and repair solutions through simulation debugging before issuing them to the on-site system for execution. A built-in maintenance knowledge base automatically associates fault types with historical repair cases, providing step-by-step repair guidance, reducing fault handling time by more than 40%.

[0039] The following two examples further illustrate the specific implementation of this system: Example 1: 300MW thermal power unit DCS intelligent transformation project (Tianjin Huadian Fuyuan Thermal Power Application Scenario) This embodiment addresses the upgrade requirements of the aging DCS system for the 300MW coal-fired power unit of Tianjin Huadian Fuyuan Thermal Power Plant. The original system relied on imported CPUs and suffered from problems such as chaotic task scheduling, closed interfaces, and slow operation and maintenance response. The solution of this invention is adopted to achieve independent and controllable upgrade, with a focus on adapting to the centralized monitoring of peak boilers and the first station of the heating network. The specific implementation process is as follows.

[0040] 1. Construction of basic hardware modules for domestically produced multi-core CPUs It adopts the Phytium FT-2000 / 4 quad-core CPU architecture, with a measured power consumption of 45W. It has achieved 18,000 hours of continuous trouble-free operation. The pin spacing and instruction set support future low-cost upgrades. It integrates a tri-mode redundancy design, with the primary and backup units synchronizing data at 1000Mbps via single-mode fiber. The arbitration unit uses an ARM Cortex-M4 core to verify the data consistency between the primary and backup units in real time. After 100 switching tests, the average switching time was 22ms, the arbitration anomaly response time was 8ms, and no data loss occurred.

[0041] Equipped with a high-speed signal acquisition module featuring a 24-bit AD converter, the sampling frequency is set to 200kHz. The module integrates dual opto-isolation circuits (5000V isolation voltage) and an aluminum alloy shielding cavity. During pulse group interference testing (±2kV, 5kHz), the sampling error remained stable at 0.04%. A single module can access 64 signals, covering boiler main steam pressure (4-20mA analog), coal feeder on / off status (DI digital), turbine speed (1kHz pulse), and RS485 bus signals from the heating network's primary station. The built-in automatic calibration circuit activates every 12 hours, calibrating with a standard signal source with an accuracy of ±0.01%, ensuring that the sampling error is always ≤±0.05%.

[0042] The underlying operating environment is built on the Kylin V10 operating system, with a system startup time of 55 seconds. 3,000 monitoring points are configured according to the 300MW scale of the unit, with 2,000 points reserved for expansion. Customized API interfaces are developed simultaneously to support the access of new modules such as carbon emission monitoring in the future.

[0043] 2. Configuration of the core module for multi-task hierarchical scheduling A five-level task scheduling system is constructed: the emergency interruption level is dedicated to responding to CPU hardware failures and power supply anomalies, with a tested response time of 9ms; the real-time control level focuses on core tasks such as boiler water level regulation and turbine speed control, with a response time of 18ms; the communication interaction level is responsible for data transmission between DCS and intelligent monitoring panels and centralized monitoring modules, with a latency of 45ms; the system supervision level checks the status of signal acquisition modules and interfaces every 45ms; and the idle-time optimization level handles non-core tasks such as daily energy consumption statistics and historical data backup, with a stable resource utilization rate of 12%.

[0044] A "core binding + dynamic migration" strategy is adopted: core 0 is bound to emergency interrupt-level tasks, core 1 is bound to real-time control-level tasks, and cores 2 and 3 are responsible for scheduling the remaining three levels of tasks. The scheduling algorithm uses an improved time-slice round-robin, with a time slice length of 2ms and a task switching overhead of 0.8ms. The single-core concurrent processing capacity reaches 850 tasks / second. A simulated load test with 1000 concurrent tasks showed no latency for real-time control-level tasks, and idle optimization-level tasks did not interfere with the core control functions.

[0045] 3. Modular functional adaptation interface module deployment It is equipped with five common interfaces: Ethernet, MODBUSTCP, OPCUA, MQTT, and DDS. A data exchange rate of up to 220MB / s was measured using dedicated testing tools. It supports parsing three mainstream data formats: SQL, JSON, and Protobuf. When connected to intelligent monitoring, centralized monitoring, or unit-level one-click start / stop modules, the module achieves hot-swapping via standardized mechanical interfaces. After insertion, it automatically triggers a parameter self-matching process: within 2 seconds, it automatically verifies the signal type, range, and communication protocol. If the verification passes, the LED indicator changes from yellow to green; if the verification fails, it lights up red and displays the fault type (e.g., signal range mismatch, protocol incompatibility) on the touchscreen.

[0046] The interface adopts a dual-redundancy design. In the event of a primary interface failure, it automatically switches to the redundant interface within 40ms. Simulating 10 communication interruptions at the primary heating network station, the switching success rate was 100%, with no data transmission interruptions. The interface's operating status is displayed in real-time on the touchscreen, including communication speed, number of connected devices, and fault records, allowing maintenance personnel to easily monitor the process.

[0047] 4. Debugging of the operation status monitoring and feedback module The system collects 30 key parameters in real time, including CPU core load, task response time, module operating temperature, and interface communication rate, with a sampling interval of 45ms. A weighted fusion algorithm is used to calculate the system health status score, with hardware parameters (CPU load, module temperature, etc.) having a weight of 0.6 and software parameters (task latency, number of failures, etc.) having a weight of 0.4. In steady-state operation, the CPU load is 40% (score 90), task response time is 18ms (score 95), module temperature is 55℃ (score 90), interface rate is 200MB / s (score 95), and there are no software failures (score 100). The final health score is calculated as follows: 0.6 × (0.25 × 90 + 0.25 × 95 + 0.25 × 90 + 0.25 × 95) + 0.4 × 100 = 94 points.

[0048] When the simulated CPU load rises to 85% (10% off the threshold), the system triggers an audible and visual alarm within 800ms and automatically initiates the fault tracing process: by using cause-effect graph analysis to correlate the overload of communication tasks in core 2, a list of processing suggestions including "migrate 30% of communication tasks to core 3" and "check the data transmission frequency of the first station of the heating network" is generated. After the maintenance personnel follow the suggestions, the system returns to normal within 10 seconds.

[0049] 5. Open development support modules and subordinate unit testing The platform adopts a dual-mode configuration of "zero-code graphical interface + code-level secondary development": Heating network engineers can complete the construction of "heating network water supply temperature optimization model" in 20 minutes by dragging and dropping data acquisition and logic judgment components. It simulates temperature fluctuations of ±5℃ in an offline simulation environment. The model parameter adjustment response is timely, and the version number 2024060101 is automatically generated after online release. It supports version rollback and comparison of modified content.

[0050] Task priority dynamic evaluation unit test: Switching to an emergency scenario during unit start-up and shutdown. =0.6, =0.3, =0.2, real-time control task (boiler ignition control) T=50ms, C=0.3, S=100 points, priority score P=0.6×50+0.3×(1-0.3)×100+0.2×100=30+21+20=71 points, significantly higher than communication interaction task (P=55 points), ensuring that core control is executed first during start-up and shutdown.

[0051] The module operates with an adaptive temperature control unit: It collects the CPU core and module temperatures in real time via a distributed temperature sensor array, with a sampling interval of 100ms. When the simulated CPU core temperature rises to 65℃, the intelligent cooling fan starts and maintains 50% speed; when it rises to 75℃, the passive heatsink's phase change layer triggers phase change cooling; when it rises to 85℃, the system automatically suspends idle optimization tasks, reducing the CPU frequency to 80% of the baseline value, and resumes normal operation after the temperature drops to 70℃ within 15 seconds. Temperature calibration is automatically initiated every 24 hours, and the sensor error is stabilized at 0.7℃ after verification with a standard temperature source.

[0052] Third-party module compatibility verification: Five desulfurization optimization third-party modules were connected, and four ran successfully. α=0.8, β=0.97, M=10, and the compatibility index C=0.8×5 / (0.97×10)=4 / 9.7≈0.412 (normalized 0.94≥0.93). The automated test recorded an average module response time of 90ms and a resource utilization rate of 18%, with no conflicts found with the existing modules.

[0053] 6. Performance Data Representation Table 1: Comparison of Task Scheduling Performance between Traditional and Inventions

[0054] Explanation: The data in Table 1 comes from actual operation statistics one month after the upgrade. Traditional systems, due to their single-level scheduling, suffer from real-time control tasks being interfered with by other tasks, resulting in response times as long as 45ms and a latency rate of 8%, posing safety hazards to core controls such as boiler water level and turbine speed. Communication and system monitoring tasks also suffer from slow response and poor stability due to resource contention. This invention, through a five-level task hierarchy and core binding strategy, reduces real-time control response time by 60%, improves communication and system monitoring response efficiency by 44% and 70% respectively, and ensures orderly and delay-free execution of idle tasks, completely solving the problems of chaotic traditional scheduling and resource waste, and providing core technical support for stable unit operation.

[0055] Example 2: New 1000MW Ultra-Supercritical Unit Project (Application Scenario of a Coastal Power Generation Base) This embodiment is a new 1000MW ultra-supercritical thermal power unit project that needs to meet the requirements of high computing power, open compatibility, and intelligent operation and maintenance. The key integration is the energy consumption optimization and remote operation and maintenance module. The specific implementation process is as follows.

[0056] 1. High-specification domestic multi-core CPU hardware foundation construction It adopts the Kunpeng 920 octa-core CPU architecture, with a measured power consumption of 65W. It has been verified to operate continuously without failure for 18,000 hours, and its instruction set is compatible with future upgrades to models with higher computing power. In the tri-mode redundancy design, the arbitration unit has been upgraded to a Cortex-M7 core, the data verification rate has been increased to 200Mbps, the average time for master-slave switchover is 20ms, and the arbitration anomaly response time is 7ms.

[0057] The high-speed signal acquisition module has a sampling frequency set to 500kHz and adopts a dual opto-isolation + electromagnetic shielding design. In the fast transient pulse interference test, the sampling error is 0.03%. A single module can access 64 signals, covering high-temperature and high-pressure signals from ultra-supercritical boilers (temperature 600℃, pressure 25MPa), with an automatic calibration error consistently maintained at 0.03%. Based on the Tongxin UOS operating system, it has a startup time of 58 seconds, supports 8000 monitoring points, and reserves 4000 points for future expansion.

[0058] 2. Multi-task hierarchical scheduling and load balancing configuration Five-level task scheduling: Emergency interruption level response 10ms, real-time control level (main steam temperature regulation) response 17ms, communication interaction level latency 40ms, system supervision level cycle 40ms, and idle optimization level resource consumption 14%. Cores 0-1 are bound to high-priority tasks, cores 2-7 schedule other tasks, time slice 1ms, switching overhead 0.7ms, single core concurrency 900 messages / second.

[0059] CPU core load balancing unit test: The load balancing degree calculation formula is L=γ∑(Ci-Cavg)² / n, where γ=1.1, Ci are 0.8, 0.7, 0.3, 0.3, 0.2, 0.2, 0.2, 0.3 respectively, Cavg=0.4, n=8. The calculated L=1.1×[(0.4²+0.3²×2+0.1²×4)] / 8=1.1×0.38 / 8≈0.052≤0.08, which is considered balanced. Simulating the load on core 0 rising to 0.9 (exceeding Cavg+0.2=0.6), the system automatically migrates 30% of idle tasks to core 4, with a migration interrupt time of 7ms. After migration, Ci=0.6, 0.7, 0.3, 0.3, 0.4, 0.2, 0.2, 0.3, L=0.03, and the load balancing degree is significantly improved.

[0060] 3. Optimized deployment of interface and monitoring modules The five interfaces have a data exchange rate of 230MB / s. When connecting to energy optimization and remote maintenance modules, the hot-swap matching time is 1.8s, and the redundancy switching is 38ms. The operation status monitoring sampling interval is 40ms, and the health score is 92 points under extreme heat conditions. The simulated module temperature is 85℃, the alarm response is 750ms, the source is traced to a cooling fan failure, and the recommended list is accurate.

[0061] 4. Open development and deep testing of dependent units The open platform connects to 8 third-party carbon emission monitoring modules, 7 of which are running successfully. α=0.9, β=0.97, M=15, C=0.9×8 / (0.97×15)=7.2 / 14.55≈0.495 (normalized 0.96≥0.93). The automated test response time is 85ms and the resource usage is 17%.

[0062] Intelligent energy consumption optimization unit: When the load is 20%, shut down 4 cores at 1.0V and 60% frequency; when the load is 80%, start all cores at 1.3V and 110% frequency; generate an energy consumption report every 5 minutes; in winter, lower the hibernation threshold to 20% to reduce energy consumption by 15%; in summer, maintain 25% to avoid frequent start-stop affecting stability.

[0063] Remote diagnostics and maintenance unit: Equipped with a 5G communication module, upload frequency of 5 seconds, and real-time display of task scheduling status on the cloud platform; simulates CPU overload fault (health score of 55), automatically pushes fault snapshot to maintenance terminal, remotely starts digital twin simulation to reproduce fault, issues task migration instructions, and shortens processing time by 45% compared to on-site maintenance.

[0064] Full-level security protection: The SM4 encryption chip has passed the tamper-proof test; the AI ​​intrusion detection has an accuracy rate of 99.92% in identifying abnormal access; the account is locked after 5 failed attempts at four-factor authentication; the blockchain log is tamper-proof and meets industrial security standards.

[0065] 5. Performance Data Representation Table 2: Load balancing and energy consumption optimization effects under different load conditions

[0066] Explanation: The data in Table 2 comes from load testing during the unit commissioning phase. Traditional systems exhibit drastic load balancing (0.18-0.30) with varying loads. Under low loads, idle cores maintain high power consumption, resulting in significant energy waste. Under high loads, single-core overload causes task latency exceeding 15%. This invention uses a load balancing algorithm to stably control the load balancing within the range of 0.04-0.08, ensuring uniform load distribution across cores. Combined with intelligent energy optimization strategies, idle cores are shut down and frequency and voltage are reduced under low loads, lowering energy consumption by 30%. Under high loads, tasks are dynamically migrated, optimizing energy consumption while maintaining performance. This addresses the pain points of traditional systems—"load imbalance and high energy consumption"—achieving a dynamic balance between performance and energy consumption, perfectly adapting to the high computing power and low energy consumption requirements of ultra-supercritical units.

[0067] Reference Figure 2 This diagram visually illustrates the core advantages of the "five-level task hierarchical scheduling" of this invention. Traditional systems employ single-level scheduling, failing to differentiate task priorities. Emergency interruption responses can take up to 30ms, real-time control responses are delayed to 45ms due to interference from other tasks, communication and monitoring tasks respond slowly due to resource contention, and idle tasks execute in an unordered manner. This invention, through core binding and dynamic migration strategies, assigns dedicated cores to high-priority tasks, reducing emergency interruption and real-time control responses by 67% and 60% respectively, improving communication and monitoring response efficiency by 44% and 70%, and enabling orderly scheduling of idle tasks. This completely solves the pain points of traditional scheduling—"chaotic and disordered, with delayed core tasks"—providing millisecond-level response guarantees for unit core control and significantly improving operational stability.

[0068] Reference Figure 3 This diagram clearly demonstrates the effectiveness of the "CPU Core Load Balancing Algorithm" of this invention. Traditional systems exhibit drastic fluctuations in load balancing. Under low load, idle cores operate at high power consumption, resulting in a load balancing score of 0.25. Under high load, single-core overload causes the load balancing score to rise to 0.30, far exceeding the acceptable threshold of 0.08, easily leading to task delays or system lag. This invention dynamically calculates and migrates tasks using a load balancing formula, stabilizing the load balancing score between 0.04 and 0.08, even under full load. This means that CPU core resources are utilized evenly, avoiding single-core overload crashes and reducing energy waste from idle cores, achieving an optimal balance between performance and resource utilization.

[0069] Reference Figure 4This diagram visually demonstrates the core value of the "open development environment" in this invention. Traditional DCS systems are highly closed, with limited interface types and a lack of standardized design, resulting in low success rates for third-party module integration and the formation of information silos. This invention, through pre-built multi-type universal interfaces and an automated compatibility testing mechanism, achieves a compatibility index ≥0.92 for all five types of third-party modules, with the lowest being just meeting the standard for the load forecasting module and the highest for the energy consumption optimization module reaching 0.96. This means that third-party developers can quickly integrate dedicated functional modules without large-scale modifications to the existing system, effectively breaking down the closed barriers of traditional systems and providing flexible support for the rapid implementation of new smart power generation technologies and continuous iteration of system functions.

[0070] Reference Figure 5 This diagram clearly demonstrates the precision of the "three-level temperature regulation strategy" of this invention. Traditional systems often use a single fan for cooling, which can easily lead to hardware failure due to insufficient heat dissipation when the temperature exceeds 80°C. This invention initiates graded regulation based on temperature gradients: the fan starts at 50% speed at 65°C, phase-change cooling is added at 75°C, the CPU frequency is further reduced at 85°C, and low-priority tasks are suspended at 90°C while the fan speed is increased to 100%. The entire regulation process is gradual, and even when the temperature rises to 90°C, the hardware temperature can still be controlled within a safe range, avoiding module damage or system crashes due to overheating. This solves the problems of "lagging response and single regulation" in traditional cooling, significantly improving hardware stability and lifespan.

[0071] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A multi-task modular hierarchical control system based on a multi-core CPU architecture, characterized in that, include: The domestically produced multi-core CPU hardware module can be equipped with either a Phytium quad-core or a Kunpeng octa-core architecture and supports pin and instruction set compatibility upgrades. It integrates a tri-mode redundancy design, with primary / backup switching synchronized via fiber optic cable, and an arbitration unit verifies data consistency. Equipped with a data acquisition module, featuring dual opto-isolation and metal shielding, it meets IEC standards for interference resistance and supports signal access. Based on Kylin V10 or UnionTech UOS, it supports monitoring extensions; The core module for multi-task hierarchical scheduling constructs a five-level scheduling system: emergency interruption level, real-time control level, communication interaction level, system supervision level, and idle time optimization level; it adopts a core binding + dynamic load migration strategy, with emergency / real-time tasks bound to a dedicated core, and other tasks rotating according to an improved time slice. Modular functional adapter interface module, with five pre-built general interfaces, supports parsing four data formats; adapts to 12 types of functional modules, with standardized interfaces, supports hot-swapping, and automatic parameter matching; dual-channel redundancy design, with LED indicator lights to show status. The operation status monitoring and feedback module collects key parameters in real time, generates a health score using a weighted fusion algorithm, displays the data and curves on the touch screen, and supports multi-dimensional filtering; Open development support modules, dual-mode configuration platform, providing drag-and-drop components; encapsulated intelligent algorithm classification repository; supports model simulation and release, with version rollback and log recording; Third parties can integrate via the SDK and upload using custom algorithms.

2. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a dynamic task priority evaluation unit, the evaluation method of which is... Where P represents the task priority score; Weighting based on task timeliness; As a weight for resource consumption; The weights are: T = weights for the associated influence; C = weights for the remaining time before the task deadline; S = weights for the CPU resources required by the task; and S = weights for the importance of the associated devices. This evaluation allows for dynamic adjustment of the task execution order.

3. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a module operating temperature adaptive adjustment unit; an integrated distributed temperature sensor array to collect temperature data in real time; and a three-level adjustment strategy of "active heat dissipation + passive heat dissipation + task control": when the temperature is ≥65℃, the intelligent cooling fan is activated and the speed is adjusted according to the temperature gradient. When the temperature is ≥75℃, the phase change heat dissipation layer of the passive heat sink is activated; When the temperature is ≥85℃, the idle optimization level tasks are temporarily suspended, and the real-time control level tasks are assigned to the cryogenic core.

4. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a third-party module compatibility verification unit, with the verification method being C=αN / βM; where C is the compatibility index; α is the third-party module adaptation success rate; N is the total number of third-party modules connected; β is the system's inherent module adaptation coefficient; and M is the total number of system's inherent modules. It is equipped with a fully automated testing mechanism, which automatically executes "basic function testing, stress testing, and fault injection testing" after the module is connected, records the module response time, resource utilization, and number of conflicts, and generates a compatibility report containing test data and optimization suggestions.

5. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes an intelligent energy consumption optimization unit that communicates with the CPU hardware basic module and the multi-task hierarchical scheduling core module; it adopts a collaborative strategy of "dynamic voltage and frequency adjustment + core hibernation + peripheral control": when the system load is ≤25%, it automatically shuts down half of the idle CPU cores; when the load is 25%-75%, it maintains the core voltage at 1.2V. When the load is ≥75%, all cores are started; this unit generates an energy consumption analysis report every 5 minutes, compares the energy consumption of a unit task under different loads, and optimizes parameters and adjusts thresholds based on historical data.

6. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a CPU core load balancing unit, with the following control method: ;in For load balancing; This is the load correction factor; Let be the load rate of the i-th CPU core; The average load rate across all CPU cores; This represents the total number of CPU cores. Overloaded cores are identified in real time through this calculation, and the task migration process is automatically initiated: idle optimization tasks are migrated first, followed by communication and interaction tasks. Before migration, the task status is saved using data snapshot technology, and the task running status is verified after migration. For real-time control level tasks, a "pre-migration + seamless switching" mode is adopted.

7. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a data breakpoint resume unit; integrates a 16GB local high-speed cache, prioritizing data according to "real-time control data > communication interaction data > system monitoring data > idle optimization data"; when the communication link is interrupted, it automatically stores untransmitted data into the cache according to priority; after the link is restored, it starts a priority resume mechanism, prioritizing the resume transmission of real-time control data, and resuming the transmission of other data in chronological order; the unit has a built-in dual data verification mechanism: generating a CRC32 checksum before transmission and comparing the checksum after transmission; after the resume is completed, it generates an integrity report, marking missing data and retransmission status; it supports linkage with cloud storage, automatically uploading low-priority data to the cloud when the local cache is about to overflow.

8. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a customized task configuration unit; provides a 10.1-inch touch-screen visual configuration interface with preset templates for three typical scenarios: "thermal power unit control, heating network scheduling, and equipment operation and maintenance"; configuration parameters cover task priority, execution cycle, CPU core binding, and resource allocation ratio; configuration parameters are stored in non-volatile flash memory and automatically loaded after system restart; it has a built-in AI parameter recommendation engine that generates three parameter schemes for selection based on the user-input scenario type, task type, and number of devices, combined with historical best configuration data; and it also has parameter validity verification functions.

9. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a full-layer security protection unit, constructing a five-dimensional protection system of "hardware-software-data-access-logs"; at the hardware level, it adopts the national cryptographic SM4 encryption chip; at the software level, it integrates an intrusion detection model based on deep learning; at the data level, it adopts the AES-256 encryption algorithm to ensure transmission security, and the stored data adopts fragmented encryption + distributed storage; at the access level, it implements five-level permission management and adopts four-fold authentication of "account + password + fingerprint / facial biometric recognition + USBKey"; at the log level, it adopts blockchain evidence storage technology to record all operations and abnormal events.

10. The multi-task modular hierarchical control system based on a multi-core CPU architecture according to claim 1, characterized in that, It also includes a remote diagnostics and maintenance unit; equipped with a 4G / 5G / Wi-Fi tri-mode communication module, it uploads system operation data to the cloud maintenance platform in real time; maintenance personnel can log in to the platform via PC or mobile terminal to remotely view real-time data, historical trends, and fault records, and issue parameter adjustment commands; when the system triggers a serious fault, it automatically pushes a fault snapshot to the maintenance terminal, supports remotely starting fault simulation, and maintenance personnel can optimize and repair solutions through simulation debugging.

Images