Air temperature adjusting system for liquid cooling electric control room based on multi-mode switching

Abstract

The invention relates to the technical field of liquid cooling electric control temperature adjustment, and discloses an air temperature adjustment system for a liquid cooling electric control room based on multi-mode switching. The system comprises a multi-source thermal state monitoring unit, a dynamic thermal coupling evaluation unit, a cooling requirement grading unit and a multi-mode decision-making unit. The multi-source thermal state monitoring unit collects the temperature of a liquid cooling loop of a frequency converter and a fracturing rectifier, the environment temperature of an electric control room and equipment operation power data in real time. The dynamic thermal coupling evaluation unit constructs a nonlinear heat conduction model based on the collected data, and evaluates the dynamic thermal coupling effect between the two devices; the cooling demand grading unit calculates the cooling demand grades of the two devices in combination with the thermal coupling evaluation result and a preset thermal load threshold value; and the multi-mode decision-making unit generates a combined forced cooling, combined natural cooling or mixed cooling mode decision through a ternary state judgment mechanism. The system can flexibly switch cooling modes, is adaptive to different thermal load states of equipment, and gives consideration to both the cooling effect and the energy efficiency.

Description

Technical Field

[0001] This invention relates to the field of liquid-cooled electronic temperature control technology, specifically to a liquid-cooled electronically controlled air temperature control system for rooms based on multi-mode switching. Background Technology

[0002] In fields such as oil extraction and industrial manufacturing, electrical control rooms serve as critical locations for the operation of core equipment. These rooms integrate high-power electronic devices such as frequency converters and fracturing rectifiers, which continuously generate significant heat during long-term, high-load operation. As industry demands for equipment operating efficiency increase, the power density of equipment within electrical control rooms continues to rise, significantly increasing the difficulty of thermal management. Currently, commonly used air conditioning systems in electrical control rooms employ a single cooling mode, either relying entirely on forced cooling or solely on natural cooling, making it difficult to flexibly adjust according to changes in the actual heat load of the equipment.

[0003] When the frequency converter and fracturing rectifier operate simultaneously, a thermal coupling effect exists between them, and a single cooling mode cannot accurately assess this dynamic thermal interaction. Continuous forced cooling during low-load operation leads to unnecessary energy consumption and increased operating costs; relying solely on natural cooling during high-load operation fails to dissipate heat promptly, potentially causing excessively high liquid cooling circuit temperatures, affecting equipment stability, and even triggering malfunctions. Furthermore, existing temperature control systems lack comprehensive acquisition and analysis of multi-source thermal state data, failing to obtain real-time key data such as the frequency converter liquid cooling circuit temperature, fracturing rectifier liquid cooling circuit temperature, electrical control room ambient temperature, and equipment operating power. This makes it difficult to accurately determine the actual cooling needs of the equipment, further exacerbating the conflict between cooling efficiency and energy consumption.

[0004] In practical applications, different devices exhibit varying heat load characteristics. Inverters experience significant heat output fluctuations during high-frequency operation, while fracturing rectifiers have relatively stable heat output but higher overall heat output. Traditional temperature control systems are not specifically designed to address these differences; a uniform cooling strategy cannot meet the individualized cooling needs of different devices, resulting in poor cooling performance. Furthermore, due to the lack of a dynamic thermal coupling assessment mechanism, the system cannot predict the additional heat load generated by heat transfer between devices, easily leading to localized overheating or overcooling. This not only affects equipment lifespan but also reduces the operational reliability of the entire electrical control system, making it difficult to meet the high-precision, low-energy-consumption requirements of modern industry for electrical control room thermal management. Summary of the Invention

[0005] The purpose of this invention is to provide a liquid-cooled air temperature control system for electrically controlled rooms based on multi-mode switching, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides a liquid-cooled air temperature control system for electrically controlled rooms based on multi-mode switching, the system comprising:

[0007] The multi-source thermal status monitoring unit collects real-time temperature data of the inverter liquid cooling circuit, the fracturing rectifier liquid cooling circuit, the electrical control room ambient temperature, and the equipment operating power.

[0008] The dynamic thermal coupling evaluation unit, based on the liquid cooling circuit temperature data and equipment operating power data output by the multi-source thermal state monitoring unit, constructs a nonlinear heat conduction model to evaluate the dynamic thermal coupling effect between the frequency converter and the fracturing rectifier.

[0009] The cooling demand classification unit calculates the forced cooling demand level of the frequency converter and the natural cooling demand level of the fracturing rectifier based on the thermal coupling effect evaluation results output by the dynamic thermal coupling evaluation unit and the preset heat load threshold.

[0010] The multi-mode decision unit generates operation mode decision results for joint forced cooling mode, joint natural cooling mode, or hybrid cooling mode based on the forced cooling demand level and natural cooling demand level output by the cooling demand classification unit through a three-dimensional state judgment mechanism.

[0011] Preferably, the dynamic thermal coupling evaluation unit performs the following operations:

[0012] Based on the gradient changes in the temperature data of the inverter liquid cooling circuit and the temperature data of the fracturing rectifier liquid cooling circuit, a nonlinear heat conduction model describing the dual-circuit heat transfer path is established.

[0013] The conduction interference coefficient of the heat generated by the frequency converter on the liquid cooling circuit of the fracturing rectifier was calculated by numerical simulation.

[0014] When the conducted interference coefficient exceeds a predetermined threshold, an abnormal thermal coupling flag is output to the cooling demand classification unit.

[0015] Preferably, the cooling demand grading unit performs the following operations:

[0016] The abnormal thermal coupling flags output by the dynamic thermal coupling assessment unit are parsed, and the thermal load urgency assessment algorithm is activated.

[0017] Based on the weighted difference between the temperature change rate of the inverter liquid cooling circuit and the temperature change rate of the fracturing rectifier liquid cooling circuit, a quantitative value for the forced cooling demand level is generated.

[0018] A quantitative value for the natural cooling requirement level is generated based on the steady-state duration of the liquid cooling circuit of the fracturing rectifier.

[0019] Preferably, the multi-mode decision-making unit performs the following operations:

[0020] When the quantified value of the forced cooling demand level reaches the first threshold and the quantified value of the natural cooling demand level reaches the second threshold, a joint forced cooling mode instruction is generated.

[0021] When the quantized value of the forced cooling demand level is lower than the first threshold and the quantized value of the natural cooling demand level is lower than the second threshold, a joint natural cooling mode instruction is generated.

[0022] When there is a threshold difference between the quantified value of the forced cooling demand level and the quantified value of the natural cooling demand level, a mixed cooling mode instruction of forced cooling of the frequency converter and natural cooling of the fracturing rectifier is generated.

[0023] Preferably, the system further includes:

[0024] The waste heat recovery control unit captures the operation mode decision results output by the multi-mode decision unit in real time. When the decision result is a mixed cooling mode, the waste heat capture mechanism of the fracturing rectifier is activated.

[0025] The waste heat conversion execution unit converts the waste heat energy value captured by the waste heat recovery control unit into a PTC heating power compensation value.

[0026] Preferably, the waste heat recovery control unit performs the following operations:

[0027] In the hybrid cooling mode, a mapping table of the temperature of the liquid cooling circuit of the fracturing rectifier and the waste heat conversion efficiency is established;

[0028] The energy flow threshold of the waste heat capture mechanism is dynamically adjusted according to the real-time temperature requirements of the inverter's liquid cooling circuit.

[0029] Preferably, the system further includes:

[0030] The temperature decoupling execution unit receives the operation mode decision results output by the multi-mode decision unit and performs hot and cold aisle separation processing on the ambient temperature data of the power control room.

[0031] The adaptive compensation unit generates an ambient temperature compensation coefficient based on the thermal channel data separated by the temperature decoupling execution unit and feeds it back to the multi-source thermal state monitoring unit.

[0032] Preferably, the temperature decoupling execution unit performs the following operations:

[0033] A dual-channel convolutional network was used to extract the spatial distribution features and temporal fluctuation features of ambient temperature data.

[0034] By fusing spatial distribution features and temporal fluctuation features through feature recombination operations, the decoupled cold aisle control parameters and hot aisle control parameters are output.

[0035] Preferably, the adaptive compensation unit performs the following operations:

[0036] The adjustment step size of the ambient temperature compensation coefficient is calculated based on the gradient change of the thermal channel control parameters.

[0037] The ambient temperature monitoring threshold is updated by superimposing the adjustment step size with the reference temperature data collected by the multi-source thermal state monitoring unit.

[0038] Preferably, the system further includes:

[0039] The mode switching optimization unit continuously receives the PTC heating power compensation value output by the waste heat conversion execution unit, the ambient temperature compensation coefficient output by the adaptive compensation unit, and the decision log of the multi-mode decision unit.

[0040] The dynamic weight allocation unit generates mode switching threshold adjustment parameters based on the historical decision dataset provided by the mode switching optimization unit and feeds them back to the cooling demand grading unit and the multi-mode decision unit.

[0041] Compared with the prior art, the beneficial effects of the present invention are:

[0042] By collecting real-time temperature data of the inverter liquid cooling circuit, the fracturing rectifier liquid cooling circuit, the ambient temperature of the electrical control room, and the operating power of the equipment through the multi-source thermal state monitoring unit, comprehensive control of key thermal state information in the electrical control room is achieved. This breaks through the limitations of the traditional temperature control system's single and one-sided data collection and can provide comprehensive and accurate basic information for the formulation of subsequent cooling strategies.

[0043] The dynamic thermal coupling assessment unit constructs a nonlinear heat conduction model based on collected liquid cooling circuit temperature data and equipment operating power data, enabling accurate assessment of the dynamic thermal coupling effect between the frequency converter and the fracturing rectifier. This design effectively identifies additional heat load changes caused by heat transfer between equipment, avoiding insufficient or excessive cooling problems caused by neglecting thermal coupling effects in traditional systems. This makes the assessment of equipment thermal status more consistent with actual operating conditions, providing a scientific basis for subsequent cooling demand classification.

[0044] The cooling demand grading unit calculates the forced cooling demand level for the frequency converter and the natural cooling demand level for the fracturing rectifier based on the thermal coupling effect assessment results and a preset heat load threshold, enabling differentiated judgment of the cooling needs of different equipment. Compared with the uniform cooling demand determination method of traditional systems, this unit can fully consider the different heat load characteristics of the frequency converter and the fracturing rectifier, and formulate corresponding cooling demand levels for their respective thermal states, ensuring that cooling resources can accurately match the actual needs of the equipment and avoiding resource waste or insufficient supply.

[0045] The multi-mode decision unit, through a ternary state judgment mechanism, generates a decision result for a combined forced cooling mode, a combined natural cooling mode, or a hybrid cooling mode based on the cooling demand level, enabling flexible switching between cooling modes. When the equipment is operating under high load and the thermal coupling effect is significant, the system can switch to the combined forced cooling mode to quickly reduce the temperature of the liquid cooling circuit and ensure stable equipment operation. When the equipment is operating under low load and the thermal coupling effect is weak, it can switch to the combined natural cooling mode to reduce energy consumption. When there are differences in equipment heat load and a moderate thermal coupling effect, the hybrid cooling mode can meet the different cooling needs of the frequency converter and the fracturing rectifier respectively, balancing cooling effect and energy efficiency.

[0046] Through the collaborative work of its various units, the entire system forms a complete closed loop, from data acquisition and thermal status assessment to demand classification and mode decision-making. It can adjust cooling strategies in real time according to changes in the operating status of equipment within the power control room, achieving adaptive temperature control without manual intervention. This design not only enhances the automation level of power control room thermal management but also effectively balances cooling performance with energy consumption. While ensuring equipment reliability, it reduces long-term system operating costs and is suitable for power control room thermal management needs under different load fluctuation scenarios, demonstrating broad practical application value. Attached Figure Description

[0047] Figure 1 This is a timing diagram of the liquid-cooled air temperature control system for an electrical control room based on multi-mode switching, as described in this invention.

[0048] Figure 2 The flowchart for dynamic thermal coupling evaluation;

[0049] Figure 3 A flowchart for multi-modal decision-making;

[0050] Figure 4 The flowchart for temperature decoupling and compensation;

[0051] Figure 5 The flowchart for temperature decoupling. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see Figure 1The present invention provides an air temperature control system for liquid-cooled electrical control rooms based on multi-mode switching. The system includes: a multi-source thermal state monitoring unit, a dynamic thermal coupling evaluation unit, a cooling demand classification unit, and a multi-mode decision-making unit.

[0054] The multi-source thermal status monitoring unit collects real-time temperature data from the inverter's liquid cooling circuit, the fracturing rectifier's liquid cooling circuit, the electrical control room's ambient temperature, and equipment operating power. This data is acquired via a sensor network, with sensors deployed at key nodes in the liquid cooling circuit and inside the electrical control room. The sampling frequency is set to once per second to ensure real-time performance. The dynamic thermal coupling assessment unit receives the liquid cooling circuit temperature data and equipment operating power data output from the multi-source thermal status monitoring unit. Based on this data, it constructs a nonlinear heat conduction model to assess the dynamic thermal coupling effect between the inverter and the fracturing rectifier. The cooling demand grading unit calculates the forced cooling demand level of the inverter and the natural cooling demand level of the fracturing rectifier based on the thermal coupling effect assessment results output by the dynamic thermal coupling assessment unit and a preset heat load threshold. The multi-mode decision unit generates operating mode decisions based on the forced cooling demand level and natural cooling demand level output by the cooling demand grading unit through a ternary state judgment mechanism, including a combined forced cooling mode, a combined natural cooling mode, or a hybrid cooling mode. The entire system achieves logical integration of each unit through an embedded controller, ensuring real-time data synchronization and processing efficiency.

[0055] Example 1: See Figure 2 The multi-source thermal status monitoring unit continuously collects temperature and power data through PT100 temperature sensors distributed at the inlet and outlet of the liquid cooling circuit and power sensors installed on the equipment's power module. This data is transmitted to the dynamic thermal coupling evaluation unit via fieldbus. This unit first preprocesses the input inverter liquid cooling circuit temperature sequence and fracturing rectifier liquid cooling circuit temperature sequence, including data filtering and outlier removal, to eliminate interference from measurement noise.

[0056] The core function of the dynamic thermal coupling evaluation unit is to establish a nonlinear heat conduction model. Based on the principle of energy conservation, this model treats the frequency converter and fracturing rectifier as two mutually influencing heat sources, forming a heat transfer path through metal pipes and cooling media. The model considers the combined effects of convective, conductive, and radiative heat transfer. During the modeling process, the unit first calculates the temperature gradient change between adjacent time steps, i.e., the difference between the current temperature value and the previous temperature value divided by the time interval. These gradient values ​​reflect the rate and direction of heat transfer.

[0057] Based on temperature gradient data, the model further describes the heat transfer path between the two loops. Specifically, when the temperature of the inverter liquid cooling loop rises sharply, its heat may affect the temperature distribution of the fracturing rectifier loop through the shared heat dissipation structure or close-range radiation. The model quantifies this effect by establishing the relationship between heat flux density and temperature gradient. The numerical simulation method uses the finite difference method for discrete solution, transforming the continuous heat conduction problem into a discrete grid calculation. Each grid node represents a specific location in the loop, and the heat conduction between nodes is described by difference equations.

[0058] Calculating the conducted interference coefficient is a crucial step in the evaluation process. This coefficient represents the degree to which the heat generated by the frequency converter affects the temperature of the liquid cooling circuit of the fracturing rectifier. During the calculation, the heat generation of the frequency converter at its rated power is first simulated, and the temperature response of the fracturing rectifier circuit is recorded. Then, the magnitude of the interference coefficient is obtained by comparing the actual measured temperature changes with the simulation results. The calculation process employs an iterative algorithm, continuously adjusting the model parameters to ensure that the simulation results match the actual data. When the conducted interference coefficient exceeds a predetermined threshold, it indicates that there is excessive thermal coupling between the two systems, which may affect the normal operation of the equipment. At this time, the dynamic thermal coupling evaluation unit generates an abnormal thermal coupling flag. This flag contains the specific value of the interference coefficient, the time point when the threshold is exceeded, and information about the specific affected circuit. This flag is transmitted to the cooling demand classification unit via a data interface.

[0059] Upon receiving an abnormal thermal coupling flag, the cooling demand grading unit immediately activates the heat load urgency assessment algorithm. This algorithm first parses the information in the flag to determine the specific characteristics of the thermal coupling anomaly. Then, the algorithm simultaneously monitors the temperature change rate of both the inverter's liquid cooling circuit and the fracturing rectifier's liquid cooling circuit. The temperature change rate, calculated by measuring the temperature change per unit time, reflects the severity of the heat load change.

[0060] To generate a quantifiable value for the forced cooling requirement level, the algorithm calculates a weighted difference in the rate of temperature change between the two loops. The weighting coefficients are assigned based on the operating characteristics and importance of the equipment; typically, the inverter has a higher weight because its heat load changes have a greater impact on the system. The weighted difference calculation considers the relative magnitude and direction of the temperature changes in the two loops. When the inverter loop temperature rises rapidly while the fracturing rectifier loop temperature changes more slowly, the weighted difference will be larger, indicating a higher level of forced cooling is required.

[0061] Forced cooling demand level quantification values ​​are represented using an integer scale from 0 to 10. Higher values ​​indicate more urgent cooling needs. The level classification is based on statistical analysis of long-term operational data, with each level corresponding to a temperature change rate range and a weighted difference range. For example, when the weighted difference exceeds a certain set value, the level quantification value automatically increases. The cooling demand grading unit also generates natural cooling demand level quantification values. This value is calculated based on the steady-state duration of the fracturing rectifier liquid cooling circuit. The steady-state criterion is that temperature fluctuations remain within a small range, typically set to a tolerance of ±0.5 degrees Celsius. The duration is counted from the first time the temperature enters the steady-state range until the temperature exceeds the tolerance range. Natural cooling demand level quantification values ​​also use a scale from 0 to 10; the longer the duration, the lower the level value, indicating better natural cooling effectiveness.

[0062] The cooling demand grading unit continuously updates the quantified values ​​for both levels and periodically outputs them to the multi-mode decision unit. These quantified values ​​provide a quantitative basis for subsequent cooling mode decisions, enabling the system to respond accurately based on actual heat load conditions. The entire assessment and grading process is automated, requiring no manual intervention, ensuring the timeliness and accuracy of the system response.

[0063] Example 2: See Figure 3 The multi-mode decision unit continuously receives quantified values ​​of the forced cooling demand level and the natural cooling demand level from the cooling demand grading unit. These quantified values ​​are transmitted in the form of digital signals, and their update frequency is synchronized with the data acquisition unit. The decision unit internally maintains a dynamic parameter table of the first and second thresholds. These thresholds are not fixed values, but fluctuate within a certain range based on the external ambient temperature and historical operating data.

[0064] When the system detects that the quantified value of the forced cooling demand level is consistently higher than the lower limit of the first threshold interval, while the quantified value of the natural cooling demand level remains at the upper limit of the second threshold interval, the decision logic triggers the generation of a joint forced cooling mode command. This command is sent to the actuator via the digital output module, simultaneously activating forced cooling equipment such as cooling tower fans, circulating water pumps, and compressor chillers. The command includes specific equipment start-up and shutdown sequences and power settings to ensure that all cooling equipment starts according to the predetermined program. Conversely, when both quantified values ​​are simultaneously lower than the lower limit of their respective threshold intervals, the system generates a joint natural cooling mode command. This command shuts down all active cooling equipment, maintaining only the operation of the basic circulating pump, relying on natural convection between outdoor air and the heat exchanger for heat dissipation. A safety check is performed before the command is executed, confirming that the equipment temperature is within a safe range before fully switching to natural cooling mode.

[0065] The most complex situation arises when there is a significant difference between the two quantified values. When the quantified value of the forced cooling demand level is consistently higher than the first threshold while the quantified value of the natural cooling demand level is lower than the second threshold, the system determines it to be a cross-threshold difference state. At this time, a hybrid cooling mode command is generated, which includes a sub-equipment control strategy: forced cooling is initiated for the frequency converter circuit, including increasing the pump frequency and starting the refrigeration unit; while the fracturing rectifier circuit is kept in a natural cooling state, with flow rate controlled only by adjusting the valve opening.

[0066] Simultaneously with the generation of the mixed cooling mode command, the waste heat recovery control unit is immediately activated. This unit captures the current operating mode decision results via a real-time data bus. When the mixed cooling mode is detected, the waste heat capture mechanism of the fracturing rectifier is initiated. This mechanism first establishes a thermodynamic model to calculate the available waste heat energy in the liquid cooling circuit of the fracturing rectifier. The regulating valve of the heat exchanger adjusts its opening degree according to the model calculation results, controlling the working fluid flow rate to capture a specific amount of waste heat.

[0067] During waste heat capture, the control system continuously monitors the temperature changes in the fracturing rectifier circuit to ensure that waste heat extraction does not affect the natural cooling effect of the circuit. The efficiency parameters of the heat exchanger are updated in real time, including heat transfer coefficient, flow rate, and temperature difference, all of which are involved in the adjustment calculations. Dynamic adjustment of the energy flow threshold is based on the real-time temperature requirements of the inverter circuit and is achieved through a closed-loop control algorithm. When the inverter circuit temperature rises, the system allows an increase in waste heat capture; when the temperature decreases, the capture amount is reduced accordingly. The waste heat conversion execution unit receives waste heat energy value data from the waste heat recovery control unit. This data is standardized and converted into a unified energy unit. The conversion module has a built-in energy mapping algorithm that maps the waste heat energy value to the power output curve of the PTC heater. The conversion process considers the current ambient temperature, the target heating temperature, and the system's energy efficiency characteristics, outputting an appropriate PTC heating power compensation value.

[0068] The PTC heating power compensation value is output to the heater control system in the form of an analog signal to achieve precise power regulation. The compensation value is calculated using an incremental algorithm, adjusting the output in real time according to changes in waste heat energy to avoid sudden power fluctuations that could impact the system. Throughout the process, energy conversion efficiency is continuously monitored, and conversion parameters are adaptively optimized based on actual operating results.

[0069] Through this implementation method, the system achieves intelligent decision-making for cooling modes and efficient energy utilization. The multi-mode decision unit ensures that the cooling method matches the actual heat load demand, while the waste heat recovery mechanism improves the overall energy efficiency of the system. All control processes are implemented digitally, with data flow and control flow tightly coupled to form a complete temperature control system.

[0070] Example 3: See Figure 4When the waste heat recovery control unit starts operating in hybrid cooling mode, it first needs to establish a precise mapping relationship. This unit retrieves historical operating data from the data storage module, which includes operating parameters of the fracturing rectifier liquid cooling circuit and corresponding waste heat recovery efficiency records for different temperature ranges. A data fitting algorithm generates a table showing the correspondence between temperature and efficiency. This mapping table is constructed using piecewise linear interpolation, dividing the temperature range into multiple intervals, each maintaining a correspondence between temperature and efficiency values. When the liquid cooling circuit temperature changes, the control unit quickly obtains the current expected recovery efficiency through a table lookup method.

[0071] The mapping relationship is updated continuously. The control unit periodically collects new operating data, and when a systematic deviation in the temperature-efficiency correspondence is detected, the mapping table correction program is automatically initiated. The correction algorithm recalculates the efficiency values ​​for each temperature range based on recent actual operating data, ensuring the mapping table maintains high accuracy. Based on the real-time temperature requirements of the inverter's liquid cooling circuit, the waste heat recovery control unit dynamically adjusts the energy flow threshold of the waste heat capture mechanism. The adjustment of the energy flow threshold is based on the temperature control requirements of the inverter circuit and is calculated using the following formula:

[0072] Φ t =K·(T) act -T set )·ΔT max

[0073] Where: Φ t T represents the current energy flow threshold, K is the proportional adjustment coefficient, and T is the energy flow threshold at the current moment. act It is the actual temperature of the inverter circuit, T set To set the temperature value, ΔT max This represents the maximum permissible temperature difference. This calculation formula ensures that the energy flow threshold is dynamically adjusted as the temperature deviation changes, allowing the system to capture higher amounts of waste heat when the actual temperature deviates significantly from the set value.

[0074] After receiving the operating mode command from the multi-mode decision-making unit, the temperature decoupling execution unit begins processing the ambient temperature data of the power control room. This unit first obtains raw temperature data from the multi-source thermal state monitoring unit, which comes from a network of temperature sensors located in different locations within the power control room. Data processing employs a region-based partitioning method, dividing the internal space of the power control room into several virtual grids, each corresponding to a temperature acquisition point.

[0075] The hot and cold aisle separation process employs a feature extraction algorithm to first identify the spatial distribution pattern of temperature data. The algorithm analyzes the correlation between data from various sensors, grouping areas with high correlation into the same aisle. For the cold aisle, it focuses on identifying areas with lower temperatures and smaller fluctuations; for the hot aisle, it focuses on areas with higher temperatures and more drastic temperature changes. The separation process also considers equipment layout and airflow organization characteristics to ensure that the division results conform to the actual temperature distribution patterns. After processing, the temperature decoupling execution unit outputs two independent sets of control parameters. The cold aisle control parameters include the target temperature range, allowable fluctuation range, and adjustment response time; the hot aisle control parameters include information such as heat concentration, temperature change rate, and heat load distribution. These parameters provide detailed information for subsequent temperature control.

[0076] The adaptive compensation unit receives the thermal channel data output from the temperature decoupling execution unit and begins calculating the ambient temperature compensation coefficient. The calculation process first analyzes the statistical characteristics of the thermal channel data, including the mean, variance, and trend. Based on the gradient change of the thermal channel control parameters, the compensation unit uses a sliding window algorithm to calculate the temperature change rate over the most recent period. The rate of change is calculated based on time series analysis, obtaining the magnitude of change by comparing data differences between adjacent time points. The adjustment step size of the ambient temperature compensation coefficient is determined through an optimization algorithm that considers multiple factors, including the current value of the thermal channel temperature, historical trends, and external environmental conditions. The adjustment step size calculation employs a gradual strategy, using a larger step size when the temperature change is large and a smaller step size when the change is small, ensuring the smoothness of the adjustment process.

[0077] The calculated adjustment step size is superimposed on the reference temperature data collected by the multi-source thermal state monitoring unit. The reference temperature data comes from the calibrated main temperature sensor and represents the baseline temperature level of the electrical control room. The superposition operation uses a weighted average method, combining the adjustment step size and the reference data in a certain proportion to generate an updated ambient temperature monitoring threshold.

[0078] The updated threshold is fed back to the multi-source thermal state monitoring unit to adjust the temperature monitoring sensitivity. The monitoring unit reconfigures the sensor's acquisition parameters, including sampling frequency, filtering parameters, and alarm thresholds, based on the new threshold range. This process forms a closed-loop control system, enabling temperature monitoring to adaptively follow environmental changes.

[0079] Throughout the implementation process, all units communicate via a high-speed data bus to ensure real-time and accurate data transmission. Control commands are generated and executed with timestamps for easy subsequent operational analysis and optimization. The system also features a comprehensive status monitoring mechanism that tracks the operational status and processing results of each unit in real time, ensuring the reliability and stability of the entire temperature control process.

[0080] Example 4: See Figure 5 After the temperature decoupling execution unit starts, it first receives raw ambient temperature data from the multi-source thermal state monitoring unit. This data comes from an array of temperature sensors deployed inside the electrical control room and on the surface of the equipment, with a sampling frequency set to once per second to ensure complete capture of the dynamic temperature characteristics. The data preprocessing stage includes signal filtering and outlier removal. A moving average filtering algorithm is used to smooth data fluctuations, while a statistical anomaly detection mechanism is set up to exclude invalid data caused by sensor malfunctions or transient interference.

[0081] The construction of a dual-channel convolutional network requires specialized configuration of network parameters. The spatial distribution feature extraction channel employs a two-dimensional convolutional kernel structure, with the kernel size determined based on the physical distance between the sensors and a stride set to a unit distance to ensure the capture of temperature gradient relationships between adjacent sensors. The output feature map of this channel reflects the temperature distribution pattern within the electrical control room, with high-value areas corresponding to heat accumulation zones and low-value areas corresponding to regions with better cooling effects. The temporal fluctuation feature extraction channel uses a one-dimensional temporal convolutional structure, with the kernel width set according to the typical period of temperature change, enabling the capture of periodic patterns and trends in temperature changes over time. The outputs of the convolutional layers in both channels are processed by activation functions to enhance the nonlinear expressive power of the features.

[0082] The feature recombination operation first aligns the output features of the two channels by padding or sampling to ensure the feature vectors have the same length. The aligned features are then fused using a join operation, concatenating the spatial and temporal feature vectors end-to-end to form a composite feature vector. This fused feature vector is then passed through a fully connected layer for dimensionality reduction. The number of nodes in the fully connected layer is optimized to reduce computational complexity while preserving key information. The final output includes two sets of data: cold aisle control parameters and hot aisle control parameters, used to guide the operation of the cooling and thermal management systems, respectively.

[0083] Cold aisle control parameters include specific indicators such as target temperature values, allowable fluctuation ranges, and adjustment response times. These parameters are dynamically generated based on feature analysis results. For example, when spatial characteristics indicate a higher temperature in a certain area, the cold aisle parameters will correspondingly increase the cooling requirements for that area. Hot aisle control parameters, on the other hand, contain information such as heat distribution characteristics, temperature change trends, and heat load intensity, which are used to guide the operation of heat recovery and heat reuse processes.

[0084] After receiving the thermal channel control parameters output by the temperature decoupling execution unit, the adaptive compensation unit begins calculating the ambient temperature compensation coefficient. The calculation process first analyzes the temporal variation characteristics of the thermal channel control parameters, and then calculates the gradient change of parameter values ​​over a recent period using a sliding window algorithm. The gradient calculation employs a difference method, comparing parameter differences at adjacent time points to obtain the rate and direction of change. Based on the magnitude and duration of the gradient change, the compensation unit uses a fuzzy logic algorithm to determine the adjustment step size. For larger gradient changes that last longer, a larger adjustment step size is used to quickly respond to environmental changes; for smaller and shorter gradient changes, a smaller adjustment step size is used to maintain system stability.

[0085] After the step size adjustment is calculated, it is superimposed on the reference temperature data collected by the multi-source thermal state monitoring unit. The reference temperature data comes from a rigorously calibrated master reference sensor and represents the baseline temperature level of the electrical control room. The superposition operation uses a weighted average method, combining the adjustment step size with the reference data at a certain ratio, with the ratio dynamically adjusted according to the current operating status. The updated ambient temperature monitoring threshold is sent to the multi-source thermal state monitoring unit for reconfiguring the parameter settings of the temperature acquisition system. Refer to Table 1, which shows the data for the temperature monitoring threshold adjustment.

[0086] Table 1: Record of Adjustment of Ambient Temperature Monitoring Thresholds

[0087]

[0088]

[0089] The multi-source thermal status monitoring unit updates its configuration parameters based on the received new threshold. The monitoring unit recalculates the alarm upper and lower limits for each sensor and adjusts the sensitivity settings for data acquisition. For areas where temperature values ​​consistently approach the new threshold, the monitoring unit automatically increases the sampling frequency to enhance monitoring. Simultaneously, the monitoring unit records historical threshold adjustments, forming a complete audit trail for subsequent analysis and optimization.

[0090] Throughout the implementation process, a complete closed-loop control mechanism was established. The temperature decoupling execution unit provides accurate environmental characteristic analysis, the adaptive compensation unit dynamically adjusts monitoring parameters based on the analysis results, and the multi-source thermal state monitoring unit executes specific monitoring tasks and feeds back the results to the upstream unit. This architecture ensures that environmental temperature monitoring can adapt to changes in operating conditions in real time, maintaining monitoring accuracy and reliability. All data processing and control command transmission are achieved through redundant communication channels, guaranteeing the system's operational stability in complex industrial environments. Data logging and status monitoring functions provide comprehensive information support for system maintenance and performance optimization.

[0091] Example 5: The mode switching optimization unit, acting as the system's data aggregation node, continuously receives real-time data streams from multiple sources. This data includes PTC heating power compensation values ​​periodically sent by the waste heat conversion execution unit (transmitted in floating-point format, reflecting the system's utilization of waste heat energy); ambient temperature compensation coefficients sent by the adaptive compensation unit at fixed time intervals, used to correct the ambient temperature monitoring benchmark; and detailed decision logs recorded by the multi-mode decision unit, containing structured data such as timestamps, trigger conditions, execution actions, and result status.

[0092] The mode switching optimization unit includes a data preprocessing module to standardize the input heterogeneous data. The PTC heating power compensation value undergoes dimensional unification and range normalization, the ambient temperature compensation coefficient is smoothed and filtered, and the decision logs are processed by a parsing engine to extract key fields and convert them into numerical features. The processed data is stored in a historical decision dataset in a time series. This dataset uses a circular buffer structure to retain complete operation records within the most recent period, while older data is archived to a long-term storage module.

[0093] The dynamic weight allocation unit periodically extracts data from the historical decision dataset of the mode switching optimization unit for analysis. The analysis process begins with feature engineering of the dataset, extracting multi-dimensional features including mode switching frequency, runtime of each cooling mode, temperature control accuracy, and energy recovery efficiency. These features, after standardization, are input into a machine learning-based analysis model for training. The model employs an offline-to-online training strategy, using historical data to train a weight prediction model, and then applying the model parameters to the real-time data stream.

[0094] During training, the model focuses on the correlation between mode switching decisions and actual operational effects. By analyzing system performance under different decision conditions in historical data, the model learns the weights of each factor's influence on system performance. Model outputs include parameters such as the first threshold offset, the second threshold offset, and the adjustment of weight coefficients in the forced cooling demand level calculation. These parameters are encapsulated into a mode switching threshold adjustment parameter data package, which is fed back to the cooling demand grading unit and the multi-mode decision unit via the data bus.

[0095] Upon receiving the adjusted parameters, the cooling demand grading unit first performs a safety check to ensure the new parameters are within acceptable limits. If the check passes, the unit internally updates the thresholds and weighting coefficients used in calculating the quantized values ​​for both the forced and natural cooling demand levels. The update process employs a gradual adjustment strategy to avoid system oscillations caused by sudden parameter changes. After parameter adjustments, the cooling demand grading unit records the adjustment log and notifies relevant units that the parameters have been updated.

[0096] The multi-mode decision-making unit also receives adjustment parameters from the dynamic weight allocation unit. These parameters are used to update the decision thresholds and judgment logic used in the ternary state judgment mechanism. The unit internally maintains a decision rule base, recalculating the triggering conditions for the joint forced cooling mode, joint natural cooling mode, and hybrid cooling mode based on the new parameter values. The decision-making unit also re-evaluates the currently executing mode based on the parameter adjustments, triggering mode switching if necessary.

[0097] The entire parameter adjustment process forms a complete closed-loop optimization loop. The mode switching optimization unit collects system operation data, the dynamic weight allocation unit analyzes the data and generates optimization parameters, and the cooling demand grading unit and multi-mode decision-making unit apply the new parameters to improve their decision-making logic. Through this mechanism, the system continuously adapts to changes in the operating environment and optimizes its mode switching strategy. All parameter adjustment operations are recorded in the system log, including adjustment time, adjustment content, adjustment reason, and adjustment effect, providing data support for subsequent analysis.

[0098] The system also establishes a security protection mechanism for parameter adjustments. When it detects that parameter adjustments may lead to a decrease in system performance or security risks, the protection mechanism automatically pauses parameter updates and reverts to the previous stable parameter settings. Simultaneously, it generates an early warning message to prompt maintenance personnel to conduct an investigation. This mechanism ensures the security and reliability of the system during its self-optimization process. The parameter adjustment effect evaluation module continuously monitors the system performance after adjustments, evaluating the effect by comparing key performance indicators before and after the adjustments. The evaluation results are fed back to the dynamic weight allocation unit to improve subsequent parameter generation strategies. The entire optimization process thus forms a continuous improvement cycle, enabling the system to continuously adapt to changes in operating conditions and maintain optimal operating status.

[0099] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0100] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A liquid-cooled air temperature control system for an electrically controlled room based on multi-mode switching, characterized in that, include: The multi-source thermal status monitoring unit collects real-time temperature data of the inverter liquid cooling circuit, the fracturing rectifier liquid cooling circuit, the electrical control room ambient temperature, and the equipment operating power. The dynamic thermal coupling evaluation unit, based on the liquid cooling circuit temperature data and equipment operating power data output by the multi-source thermal state monitoring unit, constructs a nonlinear heat conduction model to evaluate the dynamic thermal coupling effect between the frequency converter and the fracturing rectifier. The cooling demand classification unit calculates the forced cooling demand level of the frequency converter and the natural cooling demand level of the fracturing rectifier based on the thermal coupling effect evaluation results output by the dynamic thermal coupling evaluation unit and the preset heat load threshold. The multi-mode decision unit generates operation mode decision results for joint forced cooling mode, joint natural cooling mode, or hybrid cooling mode based on the forced cooling demand level and natural cooling demand level output by the cooling demand classification unit through a three-dimensional state judgment mechanism.

2. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 1, characterized in that, The dynamic thermal coupling evaluation unit performs the following operations: Based on the gradient changes in the temperature data of the inverter liquid cooling circuit and the temperature data of the fracturing rectifier liquid cooling circuit, a nonlinear heat conduction model describing the dual-circuit heat transfer path is established. The conduction interference coefficient of the heat generated by the frequency converter on the liquid cooling circuit of the fracturing rectifier was calculated by numerical simulation. When the conducted interference coefficient exceeds a predetermined threshold, an abnormal thermal coupling flag is output to the cooling demand classification unit.

3. The liquid-cooled electrical control room air temperature regulation system based on multi-mode switching according to claim 2, characterized in that, The cooling demand classification unit performs the following operations: The abnormal thermal coupling flags output by the dynamic thermal coupling assessment unit are parsed, and the thermal load urgency assessment algorithm is activated. Based on the weighted difference between the temperature change rate of the inverter liquid cooling circuit and the temperature change rate of the fracturing rectifier liquid cooling circuit, a quantitative value for the forced cooling demand level is generated. A quantitative value for the natural cooling requirement level is generated based on the steady-state duration of the liquid cooling circuit of the fracturing rectifier.

4. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 3, characterized in that, The multi-mode decision-making unit performs the following operations: When the quantified value of the forced cooling demand level reaches the first threshold and the quantified value of the natural cooling demand level reaches the second threshold, a joint forced cooling mode instruction is generated. When the quantized value of the forced cooling demand level is lower than the first threshold and the quantized value of the natural cooling demand level is lower than the second threshold, a joint natural cooling mode instruction is generated. When there is a threshold difference between the quantified value of the forced cooling demand level and the quantified value of the natural cooling demand level, a mixed cooling mode instruction of forced cooling of the frequency converter and natural cooling of the fracturing rectifier is generated.

5. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 4, characterized in that, Also includes: The waste heat recovery control unit captures the operation mode decision results output by the multi-mode decision unit in real time. When the decision result is a mixed cooling mode, the waste heat capture mechanism of the fracturing rectifier is activated. The waste heat conversion execution unit converts the waste heat energy value captured by the waste heat recovery control unit into a PTC heating power compensation value.

6. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 5, characterized in that, The waste heat recovery control unit performs the following operations: In the hybrid cooling mode, a mapping table of the temperature of the liquid cooling circuit of the fracturing rectifier and the waste heat conversion efficiency is established; The energy flow threshold of the waste heat capture mechanism is dynamically adjusted according to the real-time temperature requirements of the inverter's liquid cooling circuit.

7. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 6, characterized in that, Also includes: The temperature decoupling execution unit receives the operation mode decision results output by the multi-mode decision unit and performs hot and cold aisle separation processing on the ambient temperature data of the power control room. The adaptive compensation unit generates an ambient temperature compensation coefficient based on the thermal channel data separated by the temperature decoupling execution unit and feeds it back to the multi-source thermal state monitoring unit.

8. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 7, characterized in that, The temperature decoupling execution unit performs the following operations: A dual-channel convolutional network was used to extract the spatial distribution features and temporal fluctuation features of ambient temperature data. By fusing spatial distribution features and temporal fluctuation features through feature recombination operations, the decoupled cold aisle control parameters and hot aisle control parameters are output.

9. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 8, characterized in that, The adaptive compensation unit performs the following operations: The adjustment step size of the ambient temperature compensation coefficient is calculated based on the gradient change of the thermal channel control parameters. The ambient temperature monitoring threshold is updated by superimposing the adjustment step size with the reference temperature data collected by the multi-source thermal state monitoring unit.

10. The liquid-cooled electrical control room air temperature control system based on multi-mode switching according to claim 9, characterized in that, Also includes: The mode switching optimization unit continuously receives the PTC heating power compensation value output by the waste heat conversion execution unit, the ambient temperature compensation coefficient output by the adaptive compensation unit, and the decision log of the multi-mode decision unit. The dynamic weight allocation unit generates mode switching threshold adjustment parameters based on the historical decision dataset provided by the mode switching optimization unit and feeds them back to the cooling demand grading unit and the multi-mode decision unit.

Images