Power station big data and edge computing based server

Abstract

This invention discloses a server based on power plant big data and edge computing. The main body of the device includes a cabinet, an external compressor, a condenser, a bottom circulating fan, a sealed ventilation base, and a visible sealed door. The compressor and condenser, together with the air inlet at the top of the cabinet and the air duct at the bottom, form a unidirectional closed-loop cooling airflow, eliminating the need for graphite heat dissipation and water-cooled leakage structures. It is suitable for the dusty, wide-temperature, and high-vibration conditions of power plants. This invention features three core innovations: closed-loop cooling, edge-layered computing power scheduling, and an integrated shock-absorbing and dustproof base. The heat dissipation structure is not easily damaged, facilitating manufacturing and maintenance. It enables low-latency edge computing on-site at power plants, significantly reducing cloud transmission bandwidth usage. It can be widely applied to big data edge processing scenarios in wind power, photovoltaic, hydropower, and thermal power plants.

Description

Technical Field

[0001] This invention relates to the fields of power plant big data processing, edge computing server hardware heat dissipation, and real-time data processing at the power plant edge, and particularly to servers based on power plant big data and edge computing. Background Technology

[0002] With the continuous advancement of the construction of new power systems, wind-solar-storage integrated power stations, distributed photovoltaic power stations, river basin cascade hydropower stations, and large thermal power plants have all deployed massive amounts of sensor data acquisition terminals. A single medium-sized power station can generate up to TB-level structured and unstructured big data on equipment operating status, grid connection parameters, environmental meteorology, fault alarms, and energy consumption metering every day. Traditional centralized cloud server architectures suffer from three major drawbacks: First, all data collected at the power plant site is transmitted back to the cloud, causing bandwidth congestion and high latency. Low-latency services such as relay protection, wind turbine pitch control, and real-time control of photovoltaic inverters cannot meet millisecond-level response requirements. Second, power plant sites are characterized by dust, large temperature differences between day and night, severe vibrations during equipment startup and shutdown, and significant fluctuations in air humidity. General-purpose commercial big data servers have rudimentary heat dissipation structures, relying solely on passive ventilation from a single-sided axial fan. Dust easily enters the chassis and adheres to the surfaces of computing chips, memory, and hard drives, accelerating component aging. Alternating hot and cold temperatures cause condensation that corrodes circuit boards, and vibrations can lead to bad sectors on hard drives and poor contact of memory contacts. Third, existing heat dissipation solutions often use graphite heat sinks or external water-cooled plates. Graphite heat sinks are brittle and prone to cracking and failure under high-temperature conditions, while water-cooled structures have complex piping that is prone to leakage. This makes maintenance difficult for power plant operators and results in high manufacturing and maintenance costs. Existing equipment primarily relies on water tanks combined with graphite plate heat exchange for heat dissipation. This solution has significant limitations in its application to power plant edge scenarios: 1. Graphite heat sinks are prone to cracking and breakage under power plant operating conditions exceeding 40°C and day-night temperature differences exceeding 20°C, leading to a rapid decline in heat dissipation performance; 2. The water-cooling structure is susceptible to leakage. Liquid leakage is strictly prohibited in power plant control cabinets and distribution rooms, as it could cause short circuits, equipment burnout, or even safety accidents; 3. It only provides basic heat dissipation functions and lacks layered computing power scheduling, local data caching, on-site fault diagnosis, and multi-source power data compatible acquisition modules designed for power plant edge computing services. This prevents on-site processing of power plant big data at the edge, still requiring all raw data to be uploaded to the cloud, thus failing to alleviate transmission pressure. Meanwhile, current power plant edge computing servers generally suffer from limited functionality: they only have data storage and simple forwarding capabilities, lack hierarchical computing power scheduling logic, and cannot differentiate between real-time control data, statistical analysis data, and historical archive data for differentiated processing; at the hardware level, they lack closed-loop circulating cooling ducts, resulting in low heat dissipation efficiency due to mixed hot and cold airflow, and dust and moisture continuously intruding into the chassis; the equipment base lacks anti-vibration limiting structures, and the low-frequency vibrations generated by the start-up and shutdown of wind turbines and generator sets continuously impact the server storage units, significantly shortening the lifespan of hard drives; and the lack of visualized maintenance and rapid disassembly and maintenance structures makes disassembly and maintenance time-consuming and inefficient in the confined cabinet space of the power plant. In summary, existing commercial big data servers, conventional water-cooled servers, and graphite heat dissipation servers cannot meet the special operating conditions of power plants, such as high dust levels, wide temperature ranges, strong vibrations, and low-latency edge computing. There is an urgent need for a server that integrates closed-loop compression cooling, edge layered processing of multi-source power data in power plants, and vibration resistance, dust prevention, and moisture protection, based on power plant big data and edge computing. This would solve the technical defects of existing technologies, such as heat dissipation failure, insufficient edge computing power, poor adaptability to field conditions, and high maintenance difficulty. Summary of the Invention

[0003] In response to existing technologies, this invention provides a server based on power plant big data and edge computing, along with a layered processing method for power plant edge big data to address the dual pain points of power plant on-site edge computing hardware and data processing.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a server based on power plant big data and edge computing, comprising a chassis, a third air outlet at the top of the chassis, a compressor fixedly mounted on the outer wall of the chassis, the output end of the compressor connected to a condenser, a protective shell covering the compressor, and an array of heat dissipation vents inside the protective shell; a first air outlet at the bottom of the chassis, a fan coaxially mounted below the first air outlet, the fan being driven by a bottom rotating motor, and a protective cover mounted at the bottom of the rotating motor; a detachable ventilation pipe at the bottom of the chassis, with a second air outlet at the bottom of the ventilation pipe; a hinged door at the front of the chassis, with a viewing glass embedded in the middle of the door, and a push-pull handle fixedly mounted on the outer wall of the door; and a base fixedly mounted at the bottom of the chassis, with shock-absorbing round feet evenly distributed at the bottom of the base, and a limiting groove inside the base, which seals against the bottom of the ventilation pipe. The enclosure integrates a power station edge computing motherboard, a multi-source power data acquisition module, a hierarchical computing power scheduling and control module, and a local high-speed storage array. The hierarchical computing power scheduling and control module is communicatively connected to the compressor, rotary motor, multi-source power data acquisition module, and local high-speed storage array, respectively, and is used to execute the power station big data edge hierarchical processing logic and dynamically adjust the cooling and heat dissipation power and computing power allocation. The compressor, condenser, third air outlet, internal cavity of the enclosure, first air outlet, fan, ventilation pipe, and second air outlet form a unidirectional closed-loop cooling air duct. External air enters the enclosure through the third air outlet and exchanges heat with the condenser to form cold air. The cold air flows from top to bottom through the edge computing motherboard and completes heat exchange with the storage array. After absorbing heat, the hot air is drawn in by the fan through the first air outlet and discharged directionally from the second air outlet to the outside of the enclosure along the ventilation pipe. Furthermore, the hierarchical computing power scheduling and control module incorporates a data classification and discrimination unit, a computing power resource allocation unit, a heat dissipation power linkage adjustment unit, and a data compression and uploading unit. The data classification and discrimination unit receives raw sensor data from the power plant uploaded by the multi-source power data acquisition module and classifies it into three categories based on data latency requirements: real-time control level, statistical analysis level, and historical archiving level. The computing power resource allocation unit dynamically allocates edge computing motherboard computing power cores based on the data classification results. The heat dissipation power linkage adjustment unit synchronously adjusts the compressor output power and rotary motor speed according to the real-time computing power load. The data compression and uploading unit executes differentiated compression strategies for the three types of data, uploading only lightweight characteristic data to the power plant cloud platform. Furthermore, the protective shell is a sheet metal integral bending structure, the heat dissipation vent is a vertical louvered ventilation slot, and a buffer heat dissipation gap is reserved between the protective shell and the outer wall of the box to isolate the compressor's operating vibration and assist the refrigeration unit in heat dissipation. Furthermore, the visible glass is double-layered hollow tempered dustproof glass, and a sealing rubber strip is installed between the glass and the door. After the door is closed, the front opening of the box is sealed as a whole, preventing dust and moisture from entering the box from the gaps in the door.

[0005] Furthermore, a sealing silicone gasket is provided around the inner side of the base limiting groove. After the ventilation pipe is inserted into the limiting groove, the silicone gasket fills the gap, thereby sealing the air duct between the ventilation pipe and the base and preventing dust from the bottom from flowing back into the box. Furthermore, the shock-absorbing round foot is a nitrile rubber shock-absorbing pad with a metal counterweight core embedded inside. It is evenly distributed at the four corners of the base to absorb the low-frequency vibrations generated by the start-up and shutdown of the power station unit and wind turbine, thereby reducing the impact of vibration on the local high-speed storage array disk.

[0006] Furthermore, the protective cover is a perforated sheet metal protective cover that completely covers the outside of the rotary motor, protecting the rotary motor from bottom dust and moisture corrosion, while not obstructing the airflow of the fan.

[0007] Furthermore, the condenser is inclinedly arranged inside the third air outlet at the top of the housing, so that when cold air enters the housing, it all flows through the condenser for heat exchange, thereby improving the refrigeration and heat exchange efficiency and avoiding local mixing of cold and hot airflows. Furthermore, the surfaces of the enclosure, door, base, and protective shell are all coated with an antistatic and anti-corrosion coating, making them suitable for the high dust, high humidity, and acid and alkali corrosive gas environments of power plants.

[0008] A data tiering processing method for servers based on power plant big data and edge computing, comprising the following steps: S1. Multi-source power data acquisition: The multi-source power data acquisition module collects raw power time-series data output from power plant wind turbines, photovoltaic inverters, water turbines, transformers, environmental meteorological sensors, and relay protection devices in real time, and transmits it to the hierarchical computing power scheduling and control module. S2. Data Classification and Classification: The data classification and classification unit classifies data based on data service latency thresholds and data types: Relay protection, pitch control, and grid-connected regulation data with latency thresholds ≤10ms are classified as real-time control data; load statistics and equipment loss analysis data with latency thresholds between 10ms and 5s are classified as statistical analysis data; and equipment historical operation logs and environmental archive data with latency thresholds >5s are classified as historical archive data. S3, Dynamic computing power allocation: The computing power resource allocation unit prioritizes the high-priority computing power cores of the edge computing motherboard to process real-time control-level data, the second-priority cores to process statistical analysis-level data, and the idle low-power cores to process historical archive-level data. S4. Edge Local Computation: Performs limit violation judgment and instantaneous fault alarm calculation on real-time control level data locally, and directly outputs control commands to the power plant field execution terminal; performs hourly / daily load forecasting and equipment loss quantification calculation on statistical analysis level data; performs lossless compression and local caching on historical archive level data. S5. Heat dissipation linkage control: The heat dissipation power linkage adjustment unit collects the computing power load rate of the edge computing motherboard in real time. When the load rate is >80%, the compressor cooling power is increased and the rotary motor speed is increased to enhance the air circulation volume in the air duct. When the load rate is <30%, the cooling and fan power is reduced for energy-saving operation. S6. Differentiated Compression Upload: The data compression upload unit only uploads fault alarm characteristic values ​​for real-time control level data, uploads hourly aggregated statistical results for statistical analysis level data, and packages and compresses historical archive level data by day before uploading it on a scheduled basis. The original full data is stored locally, reducing the bandwidth usage of cloud transmission.

[0009] The technical effects and advantages of this invention are as follows: 1. The closed-loop compression refrigeration duct is suitable for extreme operating conditions in power plants, eliminating brittle graphite heat sinks and water-cooled pipes with leakage risks. The compressor and condenser achieve continuous constant temperature refrigeration. The unidirectional top-down air duct avoids mixing of hot and cold air. The third air outlet, sealed door, and base sealing limit groove are multi-layered dustproof and moisture-proof structures to prevent dust and condensation from corroding components in the power plant. The heat dissipation structure is made of sheet metal in one piece, which is simple to process and assemble, not easy to break or be damaged, and convenient for on-site operation and maintenance in the power plant. 2. Supporting the power station edge big data layered processing method, the hardware is equipped with a layered computing power scheduling and control module to realize three-level hierarchical power data, dynamic allocation of computing power, local fault calculation, and lightweight data upload. It can complete real-time control business response in milliseconds, significantly reduce the bandwidth pressure of transmission from the power station to the cloud, and solve the problems of high latency and bandwidth congestion in traditional centralized cloud architecture; the heat dissipation power is linked to the computing power load for adjustment, taking into account both heat dissipation performance and equipment operating energy consumption. 3. The integrated shock-absorbing base, combined with the hollow sealed glass door and the external compressor independent protective shell, provides multiple anti-vibration, dustproof, and anti-corrosion structures suitable for power station sites with strong vibrations during the start-up and shutdown of fans and generator sets, high dust levels, and high humidity. The visual glass allows observation of the internal computing and storage equipment operating status without opening the door. The handle, along with the quick-release base and ventilation duct structure, significantly improves maintenance efficiency in the confined power station control cabinet.

[0010] 4. The external cooling unit has an independent protective shell with louvered heat dissipation vents, which isolates the compressor vibration and ensures continuous heat dissipation of the cooling unit, avoiding the cooling system shutdown and failure under high temperature conditions, ensuring the server operates stably and uninterruptedly throughout the year, and is suitable for the 24-hour continuous data acquisition and edge computing needs of various types of power plants such as wind, solar, hydro, and thermal power plants. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a schematic diagram of the structure of the present invention.

[0013] Figure 2 This is a schematic diagram of the rear view structure of the present invention.

[0014] Figure 3 This is a schematic diagram of the structure of the present invention viewed from below.

[0015] Figure 4 This is a schematic diagram of the exploded structure of the present invention.

[0016] The attached diagrams are labeled as follows: 1. Cabinet; 2. Door; 3. Glass; 4. Handle; 5. Compressor; 6. Condenser; 7. Protective shell; 8. Heat dissipation vent; 9. First air outlet; 10. Fan; 11. Rotary motor; 12. Protective cover; 13. Ventilation duct; 14. Second air outlet; 15. Base; 16. Round foot; 17. Slot; 18. Third air outlet. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0018] Example 1 Complete assembly structure of hardware equipment, such as Figures 1-4 As shown, the present invention discloses a server based on power plant big data and edge computing. The core supporting structure is a box 1. The internal cavity of the box 1 is used to accommodate the power plant edge computing motherboard, multi-source power data acquisition module, hierarchical computing power scheduling and control module, and local high-speed storage array. All computing power and storage hardware are concentrated in the lower middle part of the box, and cold air covers all heat-generating components from top to bottom.

[0019] A long, grid-like third air outlet 18 is opened through the top surface of the housing 1. The third air outlet 18 serves as the only air intake channel for external cold air. The condenser 6 is installed at an angle on the inner side of the third air outlet. The compressor 5 is externally mounted on the right outer wall of the housing 1. The compressor 5 pipeline is sealed and connected to the condenser 6. An integrated sheet metal protective shell 7 is installed on the outside of the compressor 5. Vertical louvered heat dissipation vents 8 are evenly opened on the right side wall of the protective shell 7. A 10mm buffer gap is reserved between the protective shell 7 and the outer wall of the housing 1. On the one hand, it isolates the compressor's operating vibration. On the other hand, air flows into the gap from the heat dissipation vent 8 to continuously dissipate heat from the compressor body, avoiding overheating and shutdown of the compressor under the high temperature environment of the power station in summer.

[0020] A rectangular first air outlet 9 is opened at the center of the bottom surface of the enclosure 1. An axial fan 10 is coaxially installed directly below the first air outlet 9. The power input end of the fan 10 is connected to a rotary motor 11. A hollow protective cover 12 is fastened to the bottom of the rotary motor 11. The protective cover 12 completely covers the motor body, preventing dust and moisture from contacting the motor windings. The hollow structure does not obstruct the downward flow of air. A cylindrical ventilation pipe 13 can be detachably inserted into the bottom of the enclosure 1. A circular second air outlet 14 is opened at the bottom of the ventilation pipe 13. An integrated base 15 is fixed to the bottom of the enclosure 1. A groove 17 is opened inside the base 15. An annular sealing silicone gasket is pasted on the inside of the groove 17. After the ventilation pipe 13 is inserted into the groove 17, the silicone gasket fills the gap between the pipe and the base, realizing a complete seal at the bottom of the air duct and preventing dust and moisture from the ground from flowing back into the enclosure through the gaps in the ventilation pipe.

[0021] The four corners of the base 15 are fitted with round feet 16. The round feet 16 are integrally cast from nitrile rubber and have a solid metal counterweight core embedded inside to form a composite shock absorption structure. When the power station's wind turbine, water turbine, and generator set start and stop and generate low-frequency vibrations, the rubber layer of the round feet 16 absorbs the vibration impact, and the metal counterweight core suppresses equipment resonance, which greatly reduces the damage of vibration to the hard disk and memory gold fingers of the local high-speed storage array inside the enclosure 1 and extends the service life of the storage hardware.

[0022] The front of the enclosure 1 is hinged to the door 2. A rectangular opening in the center of door 2 houses double-layered, hollow tempered glass 3. A ring-shaped foamed rubber sealing strip is installed between the glass 3 and the door opening. When door 2 is fully closed, the sealing strip presses against the front end face of enclosure 1, sealing the front gap and preventing dust and humid air from entering from the front. A plastic push-pull handle 4 is fixed to the center of the outer side of door 2, allowing maintenance personnel to quickly open and close the door without additional tools. The double-layered, hollow glass 3 provides heat insulation, dustproofing, and visibility. Maintenance personnel can observe the operating status of the computing motherboard and hard drive indicator lights inside the enclosure without opening the door, quickly diagnosing equipment faults and reducing the risk of dust entering the enclosure due to frequent door openings.

[0023] The entire sheet metal substrate, including the enclosure 1, door 2, base 15, and protective shell 7, is coated with an anti-static epoxy anti-corrosion coating with a thickness of 80μm. It has acid and alkali resistance, anti-static properties, and moisture-proof properties, making it suitable for corrosive working conditions such as coastal wind power stations, chemical supporting thermal power plants, and high-humidity river basin hydropower stations.

[0024] The internal electrical hardware connection logic of enclosure 1: The hierarchical computing power scheduling and control module communicates bidirectionally with compressor 5, rotary motor 11, multi-source power data acquisition module and local high-speed storage array through signal lines; the hierarchical computing power scheduling and control module has built-in data classification and discrimination unit, computing power resource allocation unit, heat dissipation power linkage adjustment unit and data compression and uploading unit to realize automatic classification of power data, dynamic allocation of computing power, linkage adjustment of heat dissipation load and differentiated data compression and uploading.

[0025] The complete circulation path of the cooling and heat dissipation air duct of this equipment is as follows: ambient air from the power plant environment enters the interior of the enclosure through the third air outlet 18 at the top of the enclosure, and flows entirely through the inclined condenser 6. The compressor 5 drives the refrigerant to exchange heat inside the condenser 6, cooling the air entering the enclosure to form low-temperature cold air. The cold air sinks from top to bottom due to its own weight, completely covering the heat-generating components such as the edge computing motherboard and storage array inside the enclosure, carrying away the heat generated by the operation of the components to form high-temperature hot air. The rotating motor 11 at the bottom of the enclosure drives the fan 10 to continuously draw downwards, and the hot air converges through the first air outlet 9 into the ventilation pipe 13, and is finally discharged directionally to the outside of the equipment from the second air outlet 14 at the bottom of the ventilation pipe, forming a unidirectional closed-loop cooling air duct. There is no internal mixing of cold and hot air, and the heat dissipation efficiency is stable.

[0026] Example 2 This embodiment provides a complete data layering processing flow for the hardware device in Application Embodiment 1, adapting to multi-source power acquisition data from various types of power plants, including wind power, photovoltaic, hydropower, and thermal power. The server data layering processing method based on power plant big data and edge computing includes the following steps: S1. Multi-source power data acquisition stage: The multi-source power data acquisition module is equipped with multiple 485, Ethernet and fiber optic acquisition interfaces to collect raw power time-series data output from the power station's wind turbine pitch controller, photovoltaic inverter, transformer temperature control sensor, turbine speed governor, meteorological radiation sensor and relay protection device in real time. The acquisition frequency can reach up to 1000 times per second, and all raw data is transmitted to the hierarchical computing power scheduling and control module in real time. S2. Data Classification and Judgment Stage: The data classification and judgment unit has a built-in power business delay threshold reference table, which automatically completes the three-level classification according to the real-time requirements of data business: 1) Real-time control level: delay threshold ≤ 10ms, including relay protection trip signals, wind turbine pitch control commands, and photovoltaic grid-connected power control data; 2) Statistical analysis level: delay threshold 10ms-5s, including transformer load current, equipment temperature, hourly power generation, and loss parameters; 3) Historical archiving level: delay threshold > 5s, including equipment daily operation logs, monthly environmental meteorological data, and shutdown maintenance records; After the classification and judgment are completed, each group of data is marked with a classification label; S3, Dynamic Computing Power Allocation Phase: The computing power resource allocation unit has an 8-core edge computing motherboard computing power core built-in, and sets priority scheduling logic: high-priority cores 1-4 are exclusively allocated to real-time control-level data to ensure millisecond-level operation response; cores 5-6 are allocated to statistical analysis-level data with secondary priority; low-power idle cores 7-8 process historical archive-level lossless compression cache; when the concurrency of real-time control data surges, low-priority core computing power resources are automatically preempted to ensure that power control services are not interrupted. S4. Edge Local Computing Stage: Differentiated local computation is performed for different levels of data: Real-time control level data completes over-limit threshold comparison and instantaneous fault identification locally. Once a device fault or grid connection over-limit is detected, switch control commands are directly output to the power plant's on-site execution terminal without waiting for cloud commands; Statistical analysis level data completes hourly load forecasting, equipment heat loss quantification, and power generation efficiency calculation based on collected time-series data; Historical archive level data only undergoes lossless LZMA compression and is cached in a large capacity local storage array, without consuming real-time computing power. S5, Heat Dissipation Linkage Control Stage: The heat dissipation power linkage adjustment unit collects the overall computing power load rate of the edge computing motherboard in real time. The load rate threshold is divided into three levels: High load condition with a load rate > 80%, the compressor's cooling power is increased to 100% simultaneously, the rotary motor speed is increased to the rated maximum, and the airflow in the air duct is increased to enhance heat dissipation; Medium load condition with a load rate of 30%-80%, the compressor and fan maintain 50% of their rated power; Low load standby condition with a load rate < 30%, the compressor switches to intermittent cooling mode, the fan runs at low speed, and the overall energy consumption of the equipment is reduced. S6. Differentiated Compression and Upload Stage: The data compression and upload unit connects to the power plant's cloud platform and executes a differentiated upload strategy according to data classification and labeling: real-time control level data only uploads fault alarms and over-limit characteristic values, while the original control data is retained locally; statistical analysis level data aggregates statistical results every hour, and only the aggregated lightweight data is uploaded; historical archive level data is packaged and compressed by natural day, and uploaded to the cloud in batches during the low bandwidth period in the early morning every day; all original power acquisition data is permanently stored in the local high-speed storage array, realizing "edge-based on-site computing and lightweight feature upload to the cloud", which greatly reduces the bandwidth occupation of fiber optic transmission from the power plant to the cloud and solves the bandwidth congestion and high latency problems caused by the full data upload in traditional solutions.

[0027] Example 3 Comparative verification of heat dissipation performance under high-temperature conditions in power plants was conducted at a photovoltaic power plant with an ambient temperature of 42℃ in summer. The control group used a graphite-based ink-cooled heat dissipation server, while the test group used the equipment described in Example 1 of this invention. The equipment operated continuously at full computing power load for 72 hours, and the core test indicators are as follows: Steady-state temperature of internal chip: The control group CPU had a steady-state temperature of 87℃, and the graphite heat sink developed minor cracks after 48 hours of operation, resulting in a 12% decrease in heat dissipation efficiency; The experimental group CPU had a steady-state temperature of 62℃, and the compressor maintained stable cooling with no heat exchange structure loss. Dust intrusion: In the control group, there was no sealed air duct, and dust was visible on the circuit board inside the chassis after 72 hours; in the experimental group, there was multiple sealed air ducts, and no dust accumulation was found inside after disassembly. Safety hazards: The control group's water-cooled pipes had a risk of condensation and dripping, and liquid leakage was prohibited in the power station's power distribution room; the test group had no liquid heat exchange structure and therefore no leakage safety hazards. Computing power response latency: The control group had no edge-layered computing power scheduling, all data was uploaded to the cloud, and the average latency of control commands was 450ms; the experimental group performed local edge computing, and the average latency of real-time control commands was 6.8ms, which met the power relay protection latency standard. Maintenance difficulty: The graphite sheet replacement and water cooling pipeline leak detection procedures in the control group were complex, and each maintenance took 90 minutes; the sheet metal air duct and compressor modular disassembly and assembly in the test group took 22 minutes for each heat dissipation component maintenance.

[0028] Test results demonstrate that the device of this invention significantly outperforms existing graphite-based ink-cooled heat dissipation servers in terms of heat dissipation stability, operational safety, edge computing power adaptability, and ease of operation and maintenance under high-temperature and dusty conditions in power plants, and is fully adapted to the edge computing business needs of new power system stations.

[0029] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A server based on power plant big data and edge computing, including a housing (1), characterized in that: The top of the box (1) is provided with a third air outlet (18). A compressor (5) is installed on the outer wall of the box (1). One end of the compressor (5) is connected to a condenser (6). A protective shell (7) is provided on the outside of the compressor (5). A heat dissipation vent (8) is provided inside the protective shell (7). The bottom of the box (1) is provided with a first air outlet (9). A fan (10) is installed below the first air outlet (9). The fan (10) is driven by a bottom rotary motor (11). A protective cover (12) is provided at the bottom of the rotary motor (11). A ventilation pipe (13) is detachably installed at the bottom of the box (1). A second air outlet (14) is provided at the bottom of the ventilation pipe (13). A door (2) is installed at the front end of the box (1). A base (15) is fixed at the bottom end of the box (1). The box (1) contains a collection of... The system consists of an edge computing motherboard, a multi-source power data acquisition module, a hierarchical computing power scheduling and control module, and a local high-speed storage array. The hierarchical computing power scheduling and control module is connected to the compressor (5), the rotary motor (11), the multi-source power data acquisition module, and the local high-speed storage array. The compressor (5), the condenser (6), the third air outlet (18), the inner cavity of the box (1), the first air outlet (9), the fan (10), the ventilation pipe (13), and the second air outlet (14) form a unidirectional closed-loop cooling air duct. External air enters the box (1) through the third air outlet (18) and exchanges heat with the condenser (6) to cool down. After the cold air absorbs heat from the computing power storage hardware inside the box, the hot air is drawn out by the fan (10) and discharged from the second air outlet (14) along the ventilation pipe (13).

2. The server based on power plant big data and edge computing according to claim 1, characterized in that: The hierarchical computing power scheduling and control module has a built-in data classification and discrimination unit, computing power resource allocation unit, heat dissipation power linkage adjustment unit, and data compression and uploading unit. The data classification and discrimination unit receives the original sensor data of the power station uploaded by the multi-source power data acquisition module and classifies it into three categories according to the time delay requirements: real-time control level, statistical analysis level, and historical archiving level. The computing power resource allocation unit dynamically allocates the computing power core of the edge computing motherboard according to the data classification. The heat dissipation power linkage adjustment unit synchronously adjusts the output power of the compressor (5) and the speed of the rotating motor (11) according to the real-time computing power load. The data compression and uploading unit performs a differentiated compression strategy on the three types of data and only uploads the lightweight feature data to the power station cloud platform.

3. The server based on power plant big data and edge computing according to claim 1, characterized in that: The door (2) is fitted with double-layer hollow tempered glass (3), and a sealing rubber strip is installed between the glass (3) and the door (2). A handle (4) is fixedly installed at the front end of the door (2). After the door (2) is closed, the rubber strip seals the front opening of the box (1) to prevent dust and moisture from entering the box.

4. The server based on power plant big data and edge computing according to claim 1, characterized in that: The base (15) has shock-absorbing round feet (16) evenly installed at the bottom. The round feet (16) are nitrile rubber composite shock-absorbing pads with embedded metal counterweight cores. The base (15) has a groove (17) inside. A sealing silicone gasket is provided around the inside of the groove (17). The bottom end of the ventilation pipe (13) is inserted into the groove (17) and sealed and limited by the silicone gasket to prevent dust from flowing back into the bottom.

5. The server based on power plant big data and edge computing according to claim 1, characterized in that: The protective shell (7) is a sheet metal integral bending structure, the heat dissipation port (8) is a vertical louvered ventilation slot, and a buffer heat dissipation gap is reserved between the protective shell (7) and the outer wall of the box (1) to isolate the vibration of the compressor (5) and assist the cooling unit in heat dissipation.

6. The server based on power plant big data and edge computing according to claim 1, characterized in that: The condenser (6) is inclinedly arranged inside the third air outlet (18) at the top of the box (1). All the cold air entering the box flows through the condenser (6) to complete the heat exchange, avoiding local mixing of cold and hot air inside the box.

7. The server based on power plant big data and edge computing according to claim 1, characterized in that: The protective cover (12) is a hollow sheet metal protective cover that completely covers the outside of the rotating motor (11), protecting the motor from dust and moisture corrosion, while not obstructing the airflow of the fan (10).

8. The server based on power plant big data and edge computing according to claim 1, characterized in that: The surfaces of the enclosure (1), door (2), base (15), and protective shell (7) are all coated with antistatic and anti-corrosion coatings to adapt to the high dust, high humidity, and acid and alkali corrosion conditions of the power station site.

9. The server based on power plant big data and edge computing according to claim 1, characterized in that: The server's data tiering processing method includes the following steps: S1. Multi-source power data acquisition: The multi-source power data acquisition module collects raw power time-series data output from power plant wind turbines, photovoltaic inverters, water turbines, transformers, meteorological sensors, and relay protection devices in real time, and transmits it to the hierarchical computing power scheduling and control module. S2. Data Classification and Judgment: The data classification and judgment unit divides data into three categories based on business latency thresholds: latency ≤ 10ms is real-time control level, latency 10ms < latency ≤ 5s is statistical analysis level, and latency > 5s is historical archiving level. S3, Dynamic computing power allocation: The computing power resource allocation unit prioritizes high-priority kernels to process real-time control-level data, second-priority kernels to process statistical analysis-level data, and low-power idle kernels to process historical archive-level data. S4. Edge Local Computation: Real-time control level data completes limit violation judgment, fault alarm calculation and outputs field control commands locally; statistical analysis level data completes load forecasting and equipment loss quantification; historical archive level data performs lossless compression and local caching. S5. Heat dissipation linkage control: The heat dissipation power linkage adjustment unit collects the motherboard computing power load rate. When the load rate is >80%, the compressor and fan power are increased. When the load rate is <30%, the cooling and fan power are reduced for energy saving. S6. Differentiated compression upload: Only real-time control-level fault characteristic values, statistical analysis-level aggregation results, and historical archive-level daily packaged data are uploaded to the cloud, while the original full data is stored locally.

10. The server based on power plant big data and edge computing according to claim 9, characterized in that: In step S3, when the concurrent volume of real-time control-level data surges, the computing power resource allocation unit automatically preempts low-priority idle kernel computing power resources to ensure uninterrupted real-time power control services of the power plant; in step S6, historical archive-level data is uploaded to the power plant cloud platform in batches during the low-bandwidth period in the early morning of each day.

Images