Prefabricated modular assembly system

Abstract

The application relates to the field of power distribution systems, and discloses a prefabricated modular combined system which comprises a rectifier module, a direct-current main bus, refrigeration equipment, a temperature sensor and a master control unit. The output end of the rectifier module is connected with the direct-current main bus, and the power supply end of the main bus is electrically connected with the refrigeration equipment. The temperature sensor is arranged at the node of the heat dissipation fins of the rectifier module and the copper bus of the main bus to collect the system temperature in real time. The master control unit compares the system temperature with a preset safety temperature threshold value, calculates the cooling power compensation amount in combination with the temperature difference and the discrete difference change rate when the system temperature is out of limit, generates a target scheduling power instruction under the premise that the maximum safety operation power value is not exceeded, and sends the instruction to the refrigeration equipment to drive the refrigeration equipment to increase the cold output. The application constructs a closed loop link of temperature collection and power compensation, responds to the heating trend in advance, effectively overcomes the refrigeration thermal inertia delay, eliminates the local heat accumulation, and avoids the electrical overload of the refrigeration equipment.

Description

Technical Field

[0001] This invention relates to the field of power supply and distribution system technology, and in particular to prefabricated modular assembly systems. Background Technology

[0002] As data centers evolve towards high density and micro-modularity, a large number of power conversion devices are integrated within the systems. The AC / DC conversion process and high-current transmission generate significant heat dissipation. In existing prefabricated modular systems, power supply and distribution equipment and ambient cooling equipment often employ independent control architectures. The cooling equipment primarily relies on collecting macroscopic return or supply air temperatures within the physical channels as feedback variables, executing passive, decoupled operating logic.

[0003] With dynamic fluctuations in system load, the semiconductor switching devices of the rectifier module and the connection points of the DC main busbar will experience severe transient temperature rises. Due to the lack of direct communication and coordination between the power supply and distribution system and the cooling equipment at the underlying data level, the cooling equipment cannot perceive the actual local heating state and temperature rise rate of the core components. This discontinuity in the control link, coupled with the inherent thermal inertia and transmission delay of gas convection in physical space, causes the cooling output of the cooling equipment to lag significantly behind the actual heat generation rate of the system. When faced with a sudden increase in load, this lagging passive response mechanism can easily lead to local heat accumulation inside the cabinet that cannot be eliminated in time, thus approaching the thermal decay limit of the components.

[0004] When faced with edge conditions such as grid fluctuations or large-scale mains power drops, the existing system lacks a seamless transient hardware-level takeover mechanism between the underlying power supply topology and backup energy storage. This makes it difficult for the system to maintain stable operation of the core link under combined scenarios of power outages or thermodynamic changes, and the overall reliability is severely restricted. Summary of the Invention

[0005] The purpose of this invention is to provide a prefabricated modular assembly system that solves the problem that existing power supply modules and environmental cooling equipment are independent of each other and have a thermal inertia response delay, which makes it impossible for the system to promptly coordinate cooling and eliminate local heat accumulation based on the actual heating trend of the underlying components.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a prefabricated modular assembly system, including multiple rectifier modules, whose input terminals are connected to AC power and whose output terminals are electrically connected to a DC main busbar. The DC main busbar is built into the system, and the power supply terminal of the DC main busbar is electrically connected to a refrigeration device. Temperature sensors are deployed in the peripheral areas of the DC main busbar and the rectifier modules. The temperature sensors collect the system temperature inside the system in real time. The signal output side of the temperature sensors is communicatively connected to a main control unit, and the control output terminal of the main control unit is communicatively connected to the refrigeration device.

[0008] The main control unit has a preset safe temperature threshold. The main control unit is used to: compare the real-time collected system temperature with the safe temperature threshold; when the system temperature exceeds the safe temperature threshold, calculate the cooling power compensation amount, generate a target scheduling power command based on the cooling power compensation amount, and send the target scheduling power command to the refrigeration equipment to drive the refrigeration equipment to increase the output cooling capacity.

[0009] Preferably, the power supply terminal of the DC main busbar is also electrically connected to multiple inverter modules, and the output terminal of the inverter modules is electrically connected to an AC load. The inverter modules invert the DC power obtained from the DC main busbar into AC power and supply power to the AC load.

[0010] Preferably, the input terminal of the rectifier module is electrically connected to an automatic transfer switch, and the input terminal of the automatic transfer switch is electrically connected to two independent AC power input terminals respectively. The automatic transfer switch performs redundant power supply switching between the two independent AC power input terminals and delivers the switched AC power to the rectifier module.

[0011] In one specific embodiment, multiple rectifier modules are connected in parallel, and multiple inverter modules are connected in parallel. Both the rectifier modules and the inverter modules adopt a hot-swappable structure.

[0012] In one specific embodiment, the rectifier module, the DC main busbar, and the main control unit are integrated and packaged together inside an independent cabinet frame. The cabinet frame and the refrigeration equipment are located in the same closed micro-module physical channel, and the air supply surface of the refrigeration equipment faces the air inlet side of the cabinet frame.

[0013] Preferably, the main control unit is further configured to: acquire the basic cooling power demand fed back by the cooling equipment in real time; when the system temperature is found to be less than or equal to the safe temperature threshold, determine that the system is in a normal heating state; and at this time, directly send the acquired basic cooling power demand as the target scheduling power command to the cooling equipment.

[0014] In one specific embodiment, the logic of the main control unit in calculating the cooling power compensation amount is as follows: obtaining the temperature difference between the system temperature and the safe temperature threshold, calculating the discrete differential rate of change of the system temperature over time, converting the temperature difference using a proportional coefficient to obtain a static compensation intensity value, converting the discrete differential rate of change using a differential coefficient to obtain an advanced compensation intensity value, and algebraically adding the static compensation intensity value and the advanced compensation intensity value, using the sum as the final cooling power compensation amount.

[0015] Preferably, the main control unit also pre-stores the maximum safe operating power value that the refrigeration equipment can withstand. The specific logic of the main control unit generating the target scheduling power command is as follows: add the acquired basic refrigeration power demand to the cooling power compensation amount to obtain a temporary power demand value, compare and determine whether the temporary power demand value exceeds the maximum safe operating power value; if it does not exceed, then set the temporary power demand value as the target scheduling power command; if it exceeds, then forcibly set the maximum safe operating power value as the target scheduling power command.

[0016] In one specific embodiment, multiple temperature sensor patches are deployed at the exposed copper busbar connection nodes of the DC main bus and in the area surrounding the heat sink fins on the surface of the rectifier module.

[0017] Preferably, the energy storage access terminal of the DC main bus is also electrically connected to a backup battery array. When a large-scale power outage occurs in the external power grid at the front end of the automatic transfer switch, the reverse connection protection diode or bidirectional DC converter at the front end of the backup battery array will automatically conduct in the forward circuit, and the chemical energy stored inside the backup battery array will be converted into DC power.

[0018] The system disclosed in this invention, through the combination of the above-described structure and logical steps, exhibits the following objective technical effects:

[0019] The system constructs a hybrid AC / DC output topology and a direct-connection power supply topology for the cooling equipment, using the DC main busbar as the node. The main control unit calculates the cooling power compensation by algebraically adding the discrete differential rate of change and the temperature difference, ensuring that the generated compensation includes both the current absolute value of temperature overshoot and the temperature rise slope over time. This compensation is then sent to the cooling equipment based on an overload prevention judgment that limits the maximum safe operating power. Through the convection arrangement of the cooling equipment's air supply surface and the cabinet frame's air inlet side, the digital adjustment commands generated by the main control unit are converted into heat exchange output within the physical space of the cabinet frame. Temperature sensors are placed on the heat sink fins and exposed copper busbars, ensuring that data acquisition is based on specific spatial heat sources. The backup battery array and self-conducting physical circuitry create an uninterrupted DC power supply path even under AC power drops.

[0020] In summary, the present invention has at least one of the following beneficial technical effects:

[0021] 1. This invention obtains the temperature difference between the system temperature and the safe temperature threshold and the discrete differential change rate through the main control unit, and obtains the cooling power compensation amount by algebraic summation. Under the condition of not exceeding the maximum safe operating power value, a target scheduling power command is generated and sent to the refrigeration equipment. Through dual compensation calculation of static proportional and dynamic differential, the heating trend can be captured in advance according to the temperature rise slope and the cooling output can be increased, overcoming the thermal inertia delay in the physical cooling process. At the same time, a strict limiting mechanism avoids electrical overload of the refrigeration equipment due to over-driving.

[0022] 2. This invention directly mounts the temperature sensor onto the outer periphery of the rectifier module's heat sink fins at the connection node between the sensor and the exposed copper busbar of the DC main busbar. The cabinet frame and the cooling equipment are placed together within a closed micro-module physical channel, with the air supply surface facing the air intake side. This deployment method ensures the accuracy of the underlying temperature data. At the same time, the physical direct-blowing air duct structure ensures that the increased cold air volume from the cooling equipment can directly cover and convert into forced convection heat transfer for the components inside the cabinet frame, effectively eliminating local heat accumulation.

[0023] 3. This invention electrically connects a backup battery array to the energy storage access end of the DC main bus. Utilizing the physical characteristics of the front-end reverse connection protection diode or bidirectional DC-DC converter, it automatically conducts forward when the transient voltage drop of the DC bus is caused by a power outage in the external main grid. This enables the energy storage battery to take over the power consumption requirements of the entire DC main bus without relying on complex software intervention commands, ensuring the power supply continuity and physical safety of the system under deep AC power drop conditions. Attached Figure Description

[0024] Figure 1 This is a hardware architecture topology diagram of an embodiment of the present invention;

[0025] Figure 2 This is a flowchart illustrating the process of an embodiment of the present invention;

[0026] Figure 3 This is a time series curve diagram of an embodiment of the present invention;

[0027] Figure 4 This is a step response curve diagram of an embodiment of the present invention;

[0028] Figure 5 This is a scatter plot of an embodiment of the present invention;

[0029] Figure 6 This is an oscilloscope image of the AC power supply under deep drop transients according to an embodiment of the present invention. Detailed Implementation

[0030] See attached document Figure 1 In this embodiment, the prefabricated modular system includes a rectifier module, a DC main bus, a temperature sensor, and a main control unit. The input of the rectifier module is connected to AC power, and its output is electrically connected to the DC main bus. The DC main bus is built into the prefabricated modular system, serving as the core power collection and distribution carrier within the system. Those skilled in the art can select the appropriate circuit structure based on actual power conversion requirements regarding the topology of the AC-to-DC power devices and the switching control principle within the rectifier module.

[0031] In a preferred embodiment, the power supply end of the DC main busbar is electrically connected to a cooling device. In specific implementations, this cooling device is manifested as a data center precision air conditioner or inter-row air conditioning unit deployed within the micro-module channel. The DC main busbar directly provides the DC power required for the operation of the DC compressor and DC fan inside the aforementioned cooling device, forming a physical, direct electrical power supply path.

[0032] Temperature sensors are deployed within the prefabricated modular assembly system housing the DC main busbar. Specifically, multiple temperature sensor patches are deployed at the exposed copper busbar connection nodes of the DC main busbar and around the heat sink fins on the rectifier module surface to collect real-time system temperature data within the prefabricated modular assembly system. This system temperature physically characterizes the actual heating state of the high-power converter nodes and high-current transmission nodes in the current power supply system, providing a direct physical causal basis for subsequent cooling interventions. The temperature sensors used here can be NTC thermistors or thermocouples.

[0033] The signal output side of the temperature sensor is connected to the main control unit. The control output of the main control unit is connected to the control motherboard inside the refrigeration equipment. The main control unit is implemented at the underlying hardware circuit level as a microcontroller unit, digital signal processor, or programmable logic controller. The main control unit's storage area has preset safe temperature thresholds. These thresholds are set based on the safe junction temperature of the core power components inside the system and the aging temperature rise limit of the insulation material, and are calibrated to be between 80% and 85% of the highest withstand temperature of the components.

[0034] See attached document Figure 2 The main control unit executes closed-loop control scheduling logic based on the above hardware architecture, specifically including the following steps.

[0035] S101, the main control unit obtains the basic cooling power demand fed back by the refrigeration equipment in real time through the communication bus, and simultaneously receives the system temperature collected and summarized in real time by multiple temperature sensors. In order to ensure the logical consistency of the data in the time series, the main control unit performs timestamp alignment processing on the sampling period of the system temperature and the feedback period of the basic cooling power demand.

[0036] Considering the bottleneck effect caused by localized hotspots, the main control unit employs multi-dimensional envelope logic to process multi-source temperature data. This involves extracting the weighted maximum temperature value from multiple temperature sensors within the same sampling period, after spatial weighting, and using this value as the current system temperature for calculation. This avoids the risk of overload and missed detection due to location deviations at a single measuring point. The basic cooling power requirement refers to the minimum operating power that the cooling equipment must maintain to offset the heat generated by the external IT load.

[0037] S102, after acquiring the underlying data, the main control unit compares the real-time collected system temperature with the preset safe temperature threshold. When the system temperature is less than or equal to the safe temperature threshold, the main control unit determines that the prefabricated modular combined system is in a normal heating state. At this time, the main control unit directly sends the acquired basic cooling power demand as the target scheduling power command to the cooling equipment to maintain the normal decoupled operation of the cooling equipment.

[0038] S103, when the system temperature exceeds the safe temperature threshold, it indicates that the heat generation rate of the power supply system itself has exceeded the natural heat dissipation rate, resulting in heat accumulation. Based on the general principle of closed-loop proportional-derivative dynamic control, the main control unit triggers the dynamic power compensation calculation logic to calculate and obtain the cooling power compensation amount. In the specific software engineering implementation, the mathematical expression used by the main control unit to calculate the cooling power compensation amount is as follows:

[0039] ;

[0040] In the formula, This indicates the cooling power compensation amount calculated by the main control unit; This represents the preset proportional coefficient, which ranges from 0.5 to 2.0 and is used to determine the basic compensation response intensity. This indicates the system temperature determined in the current sampling period; Indicates the safe temperature threshold; This represents the preset differential coefficient, which is positively correlated with the physical thermal inertia of the prefabricated modular system, and its value ranges from 1.0 to 5.0. and They represent the current number. The next sampling period is the same as the previous one, i.e., the [number]th sampling period. System temperature in each sampling period; This represents the discrete time interval between two consecutive samples.

[0041] To ensure the integrity of the control algorithm and the system's resilience to downtime, the main control unit executes pre-verification logic: when it determines... When the value approaches 0 or the sampling clock stalls, the differential operation term for that cycle is forcibly discarded, and only the proportional operation term is retained to maintain basic protection.

[0042] As can be seen from the above calculation logic, the main control unit obtains the temperature difference between the system temperature and the safe temperature threshold, and calculates the discrete differential rate of change of the system temperature over time. The main control unit uses a proportional coefficient to convert the temperature difference into a static compensation strength value, and uses a differential coefficient to convert the discrete differential rate of change into a proactive compensation strength value. The main control unit algebraically adds the static compensation strength value and the proactive compensation strength value, and uses this sum as the final cooling power compensation amount. In this process, static compensation is used to eliminate steady-state thermal runaway deviations, while proactive compensation uses the temperature rise slope to suppress potential severe thermal shocks in advance, ensuring that the generation of cooling commands takes into account both the current overheating state and future heating trends.

[0043] S104, the main control unit pre-stores the maximum safe operating power value that the refrigeration equipment can withstand. This value is jointly calibrated by the maximum allowable current and heat dissipation capacity of the refrigeration equipment's compressor inverter. The main control unit adds the acquired basic refrigeration power demand to the cooling power compensation amount calculated in the above steps to obtain a temporary power demand value. The main control unit compares and determines whether this temporary power demand value exceeds the maximum safe operating power value. If it does not exceed the maximum safe operating power value, the main control unit sets the temporary power demand value as the target scheduling power command; if it exceeds the maximum safe operating power value, the main control unit forcibly sets the maximum safe operating power value as the target scheduling power command to prevent the refrigeration equipment from being overloaded due to excessive drive commands.

[0044] S105, after completing the limit protection determination, the main control unit sends the final generated target scheduling power command to the control board of the refrigeration equipment. Upon receiving the target scheduling power command, the internal frequency converter of the refrigeration equipment increases the operating frequency of the compressor or blower according to the command value, driving the refrigeration equipment to increase its output cooling capacity. The increased cold air is delivered to the physical space where the prefabricated modular assembly system is located, cooling the prefabricated modular assembly system itself, thereby establishing an effective closed-loop path for power dispatch to intervene in environmental thermodynamics.

[0045] In this embodiment, based on the aforementioned overall system architecture, the electrical nodes and physical structures within the prefabricated modular combined system are specially designed to support AC / DC hybrid output and efficient thermodynamic control.

[0046] To build a highly reliable front-end access network, the input of the rectifier module is electrically connected to an automatic transfer switch.

[0047] The automatic transfer switch has two independent AC power inputs connected to its input terminals. The automatic transfer switch is used for redundant power switching between the two independent AC power inputs and delivers the switched AC power to the rectifier module.

[0048] In practical applications, the two independent AC power inputs are provided by different physical substations or different feeder circuits, thereby eliminating the risk of single-point failure from a single AC source. For the mechanical interlocking devices, contactor drive coils, and voltage phase synchronization detection circuits inside the automatic transfer switch, those skilled in the art can select conventional CB-class or PC-class dual-power switching equipment based on the required power-off switching duration.

[0049] An inverter module is electrically connected to the power supply terminal of the DC main busbar, and the output terminal of this inverter module is electrically connected to an AC load. The inverter module is used to convert the DC power obtained from the DC main busbar into AC power and supply power to the AC load. Based on the above electrical arrangement, the prefabricated modular system, with the DC main busbar as its core hub, simultaneously possesses the capability to directly drive DC refrigeration equipment and drive conventional AC loads with a hybrid AC / DC output. The inverter module internally employs an insulated-gate bipolar transistor inverter bridge and space vector pulse width modulation generation logic.

[0050] Considering the stringent requirements for power supply continuity in critical business scenarios such as data centers, this embodiment employs an array-based design for the power conversion devices. Specifically, there are multiple rectifier modules connected in parallel, and multiple inverter modules are also connected in parallel. Both rectifier and inverter modules utilize a hot-swappable structure to enable system capacity expansion and faulty module replacement without interrupting the DC main bus power supply.

[0051] To support the aforementioned hot-swapping and parallel expansion processes, the underlying hardware circuitry and control logic must work together to achieve a smooth state transition. During physical connection, the hot-swapping structure utilizes a blind-mating connector with guide pins of varying lengths to achieve arc-prevention isolation. The longer grounding and communication pins are connected first, activating the module's internal soft-start circuit, which uses a series pre-charge resistor to limit the capacitor charging inrush current. Subsequently, the shorter power pins are connected to prevent arcing discharge between the DC main busbar and the module interface under energized conditions, thus preventing terminal burnout.

[0052] After the module is inserted and the communication handshake with the main control unit is completed, the system enters the dynamic load sharing state. Multiple modules operating in parallel are based on internal independent DC droop control logic (V-IDroopControl). By detecting their own output current and linearly adjusting the output voltage reference value, they achieve autonomous current sharing among multiple units, ensuring that no single module will fail due to overload while bearing the entire load.

[0053] See attached document Figure 1 The rectifier module, DC main busbar, and main control unit are all integrated and packaged within a single cabinet frame. The cabinet frame adopts standardized IT cabinet structural dimensions and features ventilation mesh that runs through the front and back.

[0054] As a key environmental coupling design feature of this embodiment, the aforementioned cabinet frame and the refrigeration equipment are housed within the same enclosed micro-module physical channel. This micro-module physical channel physically isolates the internal airflow from the external environment through a cold aisle sealing component or a hot aisle sealing component. The air supply surface of the refrigeration equipment faces the air intake side of the cabinet frame, allowing the increased cooling output of the refrigeration equipment to directly and physically cover and cool the physical environment surrounding the cabinet frame.

[0055] The heat accumulation elimination process within a prefabricated modular system is essentially based on the principle of forced convection heat transfer in fluid mechanics. To quantify the mapping relationship between cooling capacity transfer and temperature drop at a physical level, the physical process by which heat dissipated within the prefabricated modular system is carried away by the refrigeration equipment satisfies the following convection heat transfer balance expression:

[0056] ;

[0057] In the formula, This indicates the heat power removed from the prefabricated modular assembly system by the refrigeration equipment through airflow circulation; This indicates the specific heat capacity of air at constant pressure. Indicates the air density of the cooling airflow; This indicates the volumetric airflow rate output by the refrigeration equipment. Indicates the temperature of the hot air flowing out of the cabinet frame; This indicates the temperature of the cold air delivered by the refrigeration equipment into the physical channel of the micro-module.

[0058] In the above physical model, the selection of each parameter needs to be determined in combination with the actual working conditions of the micro-module.

[0059] As a preferred approach, within the range of standard atmospheric pressure and typical computer room operating temperature (18℃~27℃), the specific heat capacity of air at constant pressure is taken as an empirical constant of approximately 1.005 kJ / (kg·K), and the air density is taken as an empirical constant of approximately 1.2 kg / m³. 3 .

[0060] Based on this constant premise, the main control unit determines that the heat power carried away is directly linearly positively correlated with the air volume flow rate and air temperature difference of the refrigeration equipment.

[0061] In the aforementioned process, the main control unit issues a target scheduling power command to the refrigeration equipment. Its core physical purpose is to directly intervene in the variable frequency actuator of the refrigeration equipment. As the scheduling power increases, the fan speed inside the refrigeration equipment increases proportionally, thereby increasing the air volume flow rate.

[0062] Simultaneously, the increased power output of the compressor leads to a decrease in the temperature of the cold air. These multi-dimensional adjustment commands collectively enhance convective heat transfer within the physical space, effectively eliminating heat buildup on the DC main busbar and rectifier module inside the cabinet frame. Through the deployment of the above structure and the logical conversion of parameters, this solution establishes a technically supportive relationship from a hardware channel structure perspective, ensuring that power dispatch commands can inevitably be transformed into thermodynamic cooling results.

[0063] In this embodiment, to ensure that the aforementioned closed-loop collaborative scheduling logic can obtain accurate underlying physical characteristic data, a spatially targeted thermoelectric state sensing network is deployed inside the prefabricated modular assembly system.

[0064] The thermodynamic state within the prefabricated modular assembly system is not uniformly distributed, and high-power power conversion devices and high-current transmission nodes are the core heat sources causing localized heat accumulation. Based on this heat-generating spatial characteristic, this embodiment has specifically arranged the physical deployment locations of temperature sensors, and the specific sensing network construction and data processing process includes the following steps.

[0065] In the S301 rectifier module, the power semiconductor switching devices generate significant switching and conduction losses during AC / DC conversion. Heat is conducted through the internal thermally conductive substrate to the external heat sink fins. The temperature sensor is fixed to the outer periphery of the heat sink fins on the rectifier module surface using thermally conductive silicone, close to the heat outlet side at the root of the fins.

[0066] On the DC main busbar side, multiple exposed copper busbar sections have contact resistance at the assembly connection nodes, which causes localized abnormal temperature rise when a large current flows continuously through them. The temperature sensor is encapsulated in epoxy resin and directly surface-mounted on the power supply end of the DC main busbar and on the exposed areas around each busbar lap bolt.

[0067] As a preferred embodiment, the temperature sensor in this example specifically uses a high-precision PT100 platinum resistance thermometer or an NTC thermistor element.

[0068] S302, the aforementioned multi-point deployed temperature sensors, after acquiring weak analog temperature electrical signals, amplify and convert them to digital signals through a pre-amplified signal conditioning circuit. To eliminate high-frequency electromagnetic interference caused by the large-current alternating magnetic field inside the cabinet and the switching frequency, a hardware low-pass filter is embedded in the conditioning circuit.

[0069] For specific active low-pass filter circuit topologies and operational amplifier RC parameter settings, those skilled in the art can perform conventional matching according to the electromagnetic compatibility level requirements of the micro-module on-site. The digital temperature signal, after hardware anti-interference processing, is synchronously aggregated to the communication port of the main control unit through the system's independent CAN bus or RS485 industrial control bus.

[0070] Due to the complex electromagnetic environment inside the S303 cabinet, the main control unit, upon receiving multiple digital temperature signals, must perform data cleaning and fusion logic to avoid misjudgments by the control system caused by a single sensor failure or communication errors. The main control unit verifies the range rationality of the temperature data fed back from each measuring point, eliminating abrupt outliers exceeding the physical limits of the components. Subsequently, based on the thermodynamic sensitivity of different locations, the main control unit uses a spatially weighted extreme value model to extract the current effective system temperature. The specific fusion calculation formula is as follows:

[0071] ;

[0072] In the formula, Indicates the first The effective temperature value of each temperature sensor in the current sampling period. It is a positive integer; Indicates the relationship with the first The spatial thermal weighting coefficient corresponds to the physical location of each temperature sensor. This weighting coefficient ranges from 0.8 to 1.2. During initial system calibration, considering the high-speed heat dissipation of the chip directly from the rectifier module fins, resulting in a rapid heating rate, the corresponding sensor's... The value is set to 1.0~1.2; however, the thermal inertia of the edge copper busbar overlap point is relatively large and the temperature rise is relatively slow, which corresponds to the sensor's... The value is set to 0.8 to 1.0. Based on this multi-dimensional weighted calculation model, the output result no longer relies solely on the absolute extreme value of a certain part, but more realistically and comprehensively reflects the critical boundary of the core component's heat generation.

[0073] S304, while acquiring internal temperature parameters of the system, establishes a bidirectional communication connection between the control output of the main control unit and the refrigeration equipment within the physical channel of the micro-module. In specific implementation, the main control unit communicates with the control motherboard of the refrigeration equipment via an Ethernet physical interface using the Modbus TCP / IP industrial protocol.

[0074] The main control unit reads the basic cooling power demand data stored in the refrigeration equipment register according to a preset sampling period. In the low-level scheduler of the main control unit, the interrupt priority of this Ethernet communication task is set to the same response level as the CAN bus receive interrupt for acquiring the system temperature. Through this low-level triggering mechanism with strict timestamp synchronization, the main control unit ensures that the system temperature variable and the basic cooling power demand variable are precisely aligned in the time domain within the same control calculation cycle, preventing compensation calculation distortion caused by data misalignment.

[0075] Based on the aforementioned synchronous aggregation mechanism of multi-source heterogeneous data, the main control unit not only grasps the real-time thermodynamic spatial distribution of the power supply system itself, but also obtains the basic load baseline for the cooling equipment to cope with external environmental loads. Through front-end physical filtering and back-end algorithm fusion error prevention, this sensing network architecture provides reliable basic data input for the system to execute subsequent accurate dynamic power compensation algorithms and equipment overload limiting judgments.

[0076] In this embodiment, based on the underlying physical characteristic data obtained by the aforementioned deployed thermoelectric state sensing network, the main control unit executes the core closed-loop collaborative scheduling control mechanism to realize the dynamic coupling of the power supply system and the refrigeration equipment in terms of load characteristics.

[0077] The main control unit is pre-installed with a microprocessor and low-level control algorithms. Its closed-loop scheduling process includes the following steps.

[0078] S401, the main control unit compares the current system temperature with the preset safe temperature threshold. Considering the temperature sampling noise and thermal inertia delay present in the actual computer room environment, and to avoid equipment fatigue caused by frequent switching of control modes at critical points, the main control unit introduces nonlinear hysteresis dead-zone logic in this step.

[0079] Specifically, the main control unit pre-sets a dead zone hysteresis value, which ranges from 0.5℃ to 1.5℃. When the system temperature drops below the difference between the safe temperature threshold and the dead zone hysteresis value, the main control unit determines that the prefabricated modular system's own heat generation and natural heat dissipation have completely returned to thermodynamic equilibrium, and there is no risk of damaging electrical insulation or component junction temperature.

[0080] Under this operating condition, the main control unit executes a decoupled scheduling strategy, directly issuing the acquired basic cooling power demand as the target scheduling power command to the cooling equipment. Since no additional cooling compensation is added, the cooling equipment in this mode only performs basic cooling for the IT equipment operating load within the external micro-module channel, minimizing the overall operating energy consumption of the micro-module from a macro perspective.

[0081] S402, when the main control unit determines that the current system temperature exceeds the safe temperature threshold, it indicates that heat has accumulated inside the system and natural heat dissipation cannot maintain thermal balance. Based on the general principle of closed-loop proportional-derivative dynamic control, the main control unit triggers the dynamic power compensation calculation logic mentioned above, and calculates the corresponding cooling power compensation amount by calling the calculation formula.

[0082] In this mathematical model, the static proportional compensation term directly responds to the current actual steady-state temperature overshoot, and its physical purpose is to provide a base cooling factor that is proportional to the degree of overheating.

[0083] The advanced differential compensation term uses the discrete differential rate of change of temperature over time to capture the temperature rise slope. Its technical purpose is to inject cooling capacity into the refrigeration equipment in advance before the power supply system experiences a severe thermal shock due to a sudden large load current. The superposition of the dual compensation mechanism effectively overcomes the drawback of refrigerant circulation delay in the physical piping of large refrigeration equipment.

[0084] S403: After determining the cooling power compensation amount, the key engineering problem facing the system is how to ensure that the superimposed command is within the physical tolerance limits of the equipment. The main control unit calculates the algebraic sum of the basic cooling power requirement and the aforementioned cooling power compensation amount to obtain the temporary power requirement value.

[0085] The main control unit incorporates multi-dimensional amplitude limiting and reconstruction logic, including absolute amplitude and dynamic slope. One aspect involves absolute amplitude truncation:

[0086] The main control unit compares the temporary power demand value with the maximum safe operating power value pre-written in the memory. If the temporary power demand value is greater than the maximum safe operating power value, the main control unit forcibly locks the maximum safe operating power value to the reference instruction of the current cycle.

[0087] As a preferred approach, the threshold calibration of the maximum safe operating power value is based on the maximum continuous allowable drive current limit of the compressor inverter of the refrigeration equipment, which is set between 110% and 115% of the rated operating power of the inverter.

[0088] On the other hand, a rate-of-change smoothing constraint is applied:

[0089] The main control unit calculates the difference between the current cycle's baseline command and the previous cycle's historical command. If the absolute value of this difference divided by the sampling period's time interval is greater than the preset maximum allowable power change rate, the main control unit forcibly clamps the command increment to the product of this maximum allowable power change rate. The value output after the above multi-dimensional limiting verification is the final set target scheduling power command.

[0090] After completing the instruction security reconfiguration (S404), the main control unit converts the target scheduling power command into an industry-standard communication register message and sends it to the control motherboard of the refrigeration equipment. The control motherboard of the refrigeration equipment internally contains a one-dimensional mapping table between power demand and inverter frequency. Based on the received power command value, the control motherboard uses a linear interpolation algorithm to look up the corresponding one-dimensional mapping table and calculate the corresponding target operating frequency of the compressor and the target speed of the fan. To avoid pointer out-of-bounds errors due to input anomalies during table lookup, the control motherboard performs non-negative verification on the input command and limits the lookup range before performing interpolation.

[0091] For the process of driving the motor and outputting the space voltage vector pulse width modulation waveform using a specific inverter vector control algorithm on the main board of the refrigeration equipment control, those skilled in the art can perform conventional adjustments based on the parameters on the motor nameplate.

[0092] Ultimately, the actuators of the refrigeration equipment increase the compressor's power output and air volume based on the calculated frequency commands, thereby increasing the output of cold air and directing it to the physical space where the prefabricated modular assembly system is located. This execution process directly increases the intensity of forced convection heat transfer in the three-dimensional physical space, thus establishing a low-level closed-loop control link where power dispatch commands intervene in the thermodynamics of the hardware environment.

[0093] In this embodiment, to address edge cases such as occasional external hardware failures and sudden environmental changes that micro-module data centers may encounter during long-term operation, a fault-tolerant and backup redundancy mechanism is further constructed within the prefabricated modular combination system to ensure the ultimate physical security of the core power dispatch link.

[0094] When the main control unit and underlying hardware encounter an abnormal physical state, the following exception handling steps are executed.

[0095] To prevent system malfunction due to control bus failure, the S501 master control unit incorporates a heartbeat timeout timer in its underlying communication program. During normal collaborative operation, the master control unit maintains periodic heartbeat message exchanges with the refrigeration equipment's control motherboard via Ethernet. If the master control unit fails to receive a response message from the refrigeration equipment's control motherboard for several consecutive predetermined periods, the master control unit determines that the bidirectional communication link has been physically interrupted or that the network is congested.

[0096] As a preferred approach, the heartbeat timeout threshold is set to 3 to 5 standard communication cycles to filter out brief network jitter.

[0097] Under this condition, the main control unit cannot continue to issue target scheduling power commands. As a low-level hardware fault-tolerant response mechanism, the refrigeration equipment's control board, upon detecting a loss of command refresh, forcibly switches out of the closed-loop collaborative scheduling mode and automatically activates its internally integrated local open-loop temperature control logic. The refrigeration equipment directly reads the measured values ​​from its built-in local return air temperature sensor and generates a basic operating frequency command based on the independent proportional-integral-differential calculation preset within the control board. This drives the compressor to maintain basic cooling output, thereby preventing overall thermal shutdown within the micro-module's physical channel due to the loss of external bus commands.

[0098] In scenario S502, under conditions of severe mechanical failure of the refrigeration equipment or physical failure of forced convection heat transfer due to extreme external temperatures, the heat buildup inside the prefabricated modular system will irreversibly exceed the conventional safety boundaries. The main control unit extracts the current system temperature and compares it with the preset extreme critical temperature threshold in the memory.

[0099] This critical temperature threshold, in a physical sense, represents the red line temperature at which the junction temperature of a component is about to trigger thermal breakdown. It is set based on 90% to 95% of the maximum withstand temperature of the power semiconductor component. When the system temperature continues to rise and exceeds the critical temperature threshold, the main control unit triggers the active power derating protection mechanism, forcibly issuing a current limiting command to the rectifier module.

[0100] The aforementioned mechanism for suppressing power output based on temperature reversal is grounded in the general principle of energy conservation. The mathematical model used by the main control unit to calculate the target derating power is as follows:

[0101] ;

[0102] In the formula, This indicates the maximum output power of the rectifier module that has been reset by the main control unit; This indicates the rated output power of the rectifier module as specified by the manufacturer. This represents the preset linear depreciation coefficient; This indicates the aforementioned critical temperature threshold.

[0103] The main control unit limits the power conversion capability of the rectifier module based on the above algebraic equation. The value range of the linear derating factor is positively correlated with the temperature and current safe operating area curves of the power devices inside the rectifier module, and it is set to reduce the rated power by 3% to 5% for every 1°C exceeding the limit.

[0104] The core physical purpose of this operational logic is to reduce the switching and conduction losses and heat generation of power devices at the source by forcibly suppressing the activity of AC-DC conversion within the power supply system under extreme conditions where external cooling methods are completely lost. To prevent excessive attenuation of the derating command from causing power outages to critical IT loads, the main control unit introduces a lower limit clamping verification step before execution. The main control unit determines whether the calculated target derating power is lower than the preset minimum sustaining power base. If it is lower than this base, the main control unit forcibly locks the target derating power to the minimum sustaining power base.

[0105] It is important to note that if, after clamping at the minimum sustaining power level, the system temperature is still not effectively contained and triggers the set absolute meltdown threshold (set as the ultimate critical temperature threshold plus a margin of 5°C to 10°C), the main control unit will determine that the hardware faces a fire risk. At this time, the main control unit will directly issue an Emergency Power Off (EPO) hardware-level pulse, implementing physical power-off protection for the entire system by cutting off the AC contactor coil circuit of the front-end automatic transfer switch.

[0106] In addition to addressing edge-case conditions in the thermodynamic dimension, the S503 prefabricated modular system also features an electrical redundancy mechanism to handle deep AC power drops at the front end. The energy storage access point of the DC main busbar is electrically connected to a backup battery array. Under normal grid power supply conditions, the two independent mains inputs normally feed power to the system. The DC power output from the rectifier module maintains the DC main busbar voltage at a floating charge nominal value higher than the open-circuit voltage of the backup battery array, which is reverse-biased and in a silent standby state. When a large-scale power outage occurs in the external grid at the automatic transfer switch, the rectifier module stops power output due to the loss of AC input waveform. At this time, the transient voltage of the DC main busbar drops rapidly due to the continuous withdrawal of external AC loads and DC cooling equipment.

[0107] As the bus voltage potential difference flips, when the main control unit detects through its internal DC voltage sampling circuit that the voltage of the DC main bus has dropped to the critical point of the physical open-circuit voltage of the backup battery array, the reverse connection protection diodes or bidirectional DC-DC converters at the front end of the backup battery array spontaneously conduct in the forward circuit. The chemical energy stored inside the backup battery array is rapidly converted into DC electrical energy, seamlessly taking over the power consumption demand of the entire DC main bus. During this discharge process, the main control unit continuously monitors the terminal voltage of the backup battery array. When this terminal voltage drops to the sum of the products of the safe discharge cutoff voltages of individual cells, the main control unit triggers a battery disconnection isolation command, disconnecting the energy storage access circuit to prevent the lithium battery pack from collapsing due to over-discharge.

[0108] Regarding the specific lithium iron phosphate cell power balancing management algorithm used within the backup battery array, and the soft-switching control timing of the bidirectional DC-DC converter circuit during the boost discharge process, those skilled in the art can make conventional settings according to the load discharge rate requirements. The energy scheduling and high / low voltage electrical isolation processes are well-known technologies in the field and will not be elaborated upon here. Based on this redundancy mechanism of direct coupling in the DC topology, this embodiment can still maintain the power supply to the underlying control loop and core loads under extreme combined scenarios of external cold source and main power failure, relying on internal energy storage, thus achieving a high survivability rate of data center infrastructure under harsh operating conditions.

[0109] To further aid in understanding the closed-loop collaborative scheduling control mechanism and fault-tolerant redundancy mechanism based on load characteristics disclosed in this invention, a specific application example is provided below, using a typical high-density computing scenario in a micro-modular data center. By constructing a system-level dynamic simulation model of a prefabricated modular combined system, experimental verification and performance comparison of the aforementioned control logic and mathematical model of this invention are conducted.

[0110] In the verification scenario of this embodiment, the factory rated output power of the rectifier module is... The power rating is set to 100kW, and the basic cooling capacity of the refrigeration equipment is set to 10kW. The main control unit has a preset safe temperature threshold. The calibrated temperature is 75℃, which is the limiting critical temperature threshold. The calibration temperature is 90℃. For the closed-loop proportional-derivative dynamic control model, the static proportional compensation coefficient... Set to 1.2, advance differential compensation coefficient Set to 2.5, discrete time interval The setting is 1 second. The maximum safe operating power value pre-stored in the main control unit is set to 11.5kW.

[0111] See attached document Figure 3 and attached Figure 4 At the 10th minute of the simulation (t=10min), a step-like large-scale IT computing task was injected into the system, causing a sudden surge in the heat generated by the high-power converter inside the system.

[0112] Under a conventional open-loop static threshold control scheme (i.e., the comparison group shown by the dashed line in the figure), the refrigeration equipment passively responds only to the cabinet return air temperature. Figure 3 It is evident that due to the inherent thermal inertia and transmission delay of forced convection heat transfer in physical space, the current system temperature of the comparison group experienced severe overshoot, with the highest peak reaching 84.2℃, approaching the thermal decay threshold of the components. Meanwhile, Figure 4 The target cooling power of the comparison group only showed a slow upward trend, which could not effectively suppress the initial thermal shock.

[0113] After adopting the closed-loop dynamic compensation control logic disclosed in this invention (i.e., the scheme shown by the solid line in the figure), based on the aforementioned step S103, when the main control unit determines the current system temperature... Exceeding the safe temperature threshold At (75℃), immediately invoke the mathematical expression. Dynamic power compensation calculations are performed. At this abrupt change, due to the large discrete difference rate of change of temperature over time, the differential term rapidly generates a cooling power compensation amount with a significant amplitude. .

[0114] Following step S104 above, the main control unit combines the 10kW base power with... The sums yield the temporary power demand value. Since this value instantaneously exceeds the physical limits of the equipment, combined with... Figure 4It can be seen that the main control unit decisively triggered the front-end anti-collision limiting logic at t=10.3min, forcibly locking the target scheduling power command to 11.5kW (i.e., the maximum safe operating power value), and driving the compressor to operate at full load. Under the feedforward intervention based on differential lead and limiting reconstruction, the maximum overshoot of the system temperature was strictly suppressed to 76.8℃, and it quickly converged after only slightly exceeding the safe temperature threshold. The above comparative data intuitively verifies the obvious technical advantages of the mathematical model of this scheme in overcoming thermal inertia delay.

[0115] To further verify the system's fault tolerance and backup redundancy mechanisms in edge scenarios, the experiment continued to trigger extreme combined faults of external cooling failure and power grid outage.

[0116] See attached document Figure 5 and attached Figure 6 Based on the scenario set in step S502 above, at simulation time t=30min, the external cold source is completely lost (i.e., the cooling stops completely as marked in the figure).

[0117] Depend on Figure 5 It can be seen that as heat accumulation intensifies, the current system temperature... Rapid ascent. When Approaching the critical threshold At 90℃, the main control unit triggers the active power derating protection mechanism and invokes the algebraic equation. Calculate the upper limit of the output. The dotted line in the diagram indicates the upper limit of the rectified output power. Subsequently, a steep decay begins based on the linear derating factor. To prevent power interruption, at t=35min, the main control unit introduces a lower limit clamping verification step, locking the derating command at the preset minimum sustaining power base (i.e., the steady-state clamping stage marked in the figure). At this time, the system heat generation and weak natural heat dissipation reach a new dynamic equilibrium, and the system temperature... The temperature eventually stabilized at around 89.5℃. This active suppression logic based on energy conservation successfully prevented the hardware from over 90℃ and causing thermal breakdown and fire without a cooling source.

[0118] Next, based on the AC drop scenario set in step S503 above, a deep drop of the front-end AC main power supply was simulated at a simulation time of 2400 seconds (corresponding to t=40min). Observation Figure 6 At the moment the AC input waveform is lost, the bus transient voltage is affected by continuous load withdrawal, and begins to decrease capacitively from the nominal value of 380V AC mains floating charge.

[0119] When the main control unit detects that the bus voltage has dropped to 345V (the critical point of the physical open-circuit voltage of the backup battery array marked in the figure), the reverse connection protection device at the front end of the backup battery array automatically turns on in the forward direction due to the bus potential difference reversal. As shown in the shaded transient area in the figure, the backup energy storage completes seamless takeover within t+2.5ms, stabilizing the bus voltage at the 345V level. During this process, no deep voltage drop that would have caused the underlying core load to lose power occurred.

[0120] Based on the experimental verification of the above-mentioned rigorous mapping theory formula and control logic, this embodiment not only confirms the effectiveness of the closed-loop compensation formula in suppressing temperature overshoot under normal high-density computing conditions.

Claims

1. A prefabricated modular assembly system, characterized in that... The system includes multiple rectifier modules, each with an input of AC power and an output of a DC main busbar. The DC main busbar is built into the system and its power supply end is electrically connected to a refrigeration device. Temperature sensors are deployed around the DC main busbar and the rectifier modules. These temperature sensors are used to collect the system temperature in real time. The signal output side of the temperature sensors is communicatively connected to a main control unit, and the control output of the main control unit is communicatively connected to the refrigeration device. The main control unit has a preset safe temperature threshold, and the main control unit is used for: The system temperature collected in real time is compared with the safe temperature threshold. When the system temperature exceeds the safe temperature threshold, the cooling power compensation amount is calculated, a target scheduling power command is generated based on the cooling power compensation amount, and the target scheduling power command is sent to the refrigeration equipment to drive the refrigeration equipment to increase the output cooling capacity.

2. The prefabricated modular assembly system according to claim 1, characterized in that, The power supply end of the DC main bus is also electrically connected to multiple inverter modules. The output end of the inverter modules is electrically connected to an AC load. The inverter modules are used to invert the DC power obtained from the DC main bus into AC power and supply power to the AC load.

3. The prefabricated modular assembly system according to claim 1, characterized in that, The input terminal of the rectifier module is electrically connected to an automatic transfer switch. The input terminal of the automatic transfer switch is electrically connected to two independent AC power input terminals. The automatic transfer switch is used to perform redundant power supply switching between the two independent AC power input terminals and deliver the switched AC power to the rectifier module.

4. The prefabricated modular assembly system according to claim 2, characterized in that, Multiple rectifier modules are connected in parallel, and multiple inverter modules are connected in parallel. Both the rectifier modules and the inverter modules adopt a hot-swappable structure.

5. The prefabricated modular assembly system according to claim 1, characterized in that, The rectifier module, the DC main busbar, and the main control unit are all integrated and packaged inside an independent cabinet frame. The cabinet frame and the refrigeration equipment are located in the same closed micro-module physical channel, and the air supply surface of the refrigeration equipment faces the air inlet side of the cabinet frame.

6. The prefabricated modular assembly system according to claim 1, characterized in that, The main control unit is also used for: The system obtains the basic cooling power demand fed back by the cooling equipment in real time. When the system temperature is found to be less than or equal to the safe temperature threshold, it is determined that the system is in a normal heating state. At this time, the obtained basic cooling power demand is directly sent to the cooling equipment as the target scheduling power command.

7. The prefabricated modular assembly system according to claim 1, characterized in that, The logic for the main control unit to calculate the cooling power compensation amount is as follows: The temperature difference between the system temperature and the safe temperature threshold is obtained, and the discrete differential rate of change of the system temperature over time is calculated. The static compensation intensity value is obtained by converting the temperature difference using a proportional coefficient, and the advanced compensation intensity value is obtained by converting the discrete differential rate of change using a differential coefficient. The static compensation intensity value and the advanced compensation intensity value are algebraically added together, and the sum is used as the final cooling power compensation amount.

8. The prefabricated modular assembly system according to claim 6, characterized in that, The main control unit also pre-stores the maximum safe operating power value that the refrigeration equipment can withstand. The specific logic for the main control unit to generate the target scheduling power command is as follows: The obtained basic cooling power requirement is added to the cooling power compensation amount to obtain the temporary power requirement value. The temporary power requirement value is then compared and it is determined whether the temporary power requirement value exceeds the maximum safe operating power value. If the target power requirement is not exceeded, then the temporary power requirement value is set to the target scheduling power command. If the maximum safe operating power value is exceeded, the maximum safe operating power value will be forcibly set to the target scheduling power command.

9. The prefabricated modular assembly system according to claim 2, characterized in that, The multiple temperature sensor patches are deployed at the exposed copper busbar connection nodes of the DC main bus and around the heat sink fins on the surface of the rectifier module.

10. The prefabricated modular assembly system according to claim 3, characterized in that, The energy storage access terminal of the DC main bus is also electrically connected to a backup battery array. When a large-scale power outage occurs in the external power grid at the front end of the automatic transfer switch, the reverse connection protection diode or bidirectional DC converter at the front end of the backup battery array will automatically conduct in the forward circuit, and the chemical energy stored inside the backup battery array will be converted into DC power.

Images