Industrial control system of micro-module moving ring system

Abstract

The invention relates to the technical field of industrial control systems, and discloses an industrial control system of a micro-module moving ring system, which comprises the following steps of: acquiring an active power signal by using an edge controller, identifying a load attribute by calculating a signal fluctuation variance value, and triggering feed-forward control when a power change rate meets a gating condition, according to the method, the false disturbance of the non-thermal electrical load is effectively eliminated, the physical time difference of energy input and medium response is eliminated, and the performance of the system is improved. And precise dynamic synchronization of refrigeration supply and actual thermal load is realized.

Description

Technical Field

[0001] This invention relates to an industrial control system for a micro-module dynamic environmental system, belonging to the field of industrial control system technology. Background Technology

[0002] Current micro-modular data centers employ a hot-and-cold aisle isolation architecture to construct a closed thermodynamic environment. The integrated power supply and cooling subsystems must maintain dynamic thermal balance. Existing industrial control systems typically use deviation feedback-based regulation strategies. This involves collecting air temperature signals from temperature sensors placed within the aisles, inputting them as process variables to a PID controller. Based on the deviation between the setpoint and the process variable, control commands are calculated to drive the air conditioning compressor or fan actuator to adjust the cooling capacity or airflow output. This type of control logic relies on air temperature feedback to eliminate existing temperature deviations. Current technological advancements in this field largely focus on physical space intensification and ease of equipment installation. For example, Chinese invention patent application CN120321903A discloses a door box device and a data center micro-module channel device. By prefabricating and integrating the dynamic environment control unit inside the door, it solves the pain points of rack space occupation and complicated on-site wiring. However, such simple structural improvements do not touch the core of the control algorithm. The supporting control strategy is still limited by the traditional feedback adjustment framework and lacks the ability to deeply identify the load attributes. When faced with pulsed heat loads caused by high-frequency quantitative trading, physical structure optimization alone cannot eliminate the time difference between energy input and temperature response. The system still faces the fundamental problems of response lag and cooling capacity mismatch.

[0003] In high-frequency computing and quantitative trading applications, IT equipment acts as a highly dynamic electrothermal conversion source. The input power surges dramatically within milliseconds and is converted into heat energy, which is conducted from the chip source to the heat sink and then migrates to the temperature sensor via air convection. This physical transmission process has inherent thermal inertia and time constants. There is a physical timing mismatch between the instantaneous nature of the electrical input and the hysteresis of the medium response. In feedback control mode, the control system needs to wait for the air temperature to change before initiating the adjustment action, resulting in a lag in heat release on the cooling response time axis. By the time the control system detects the high temperature and increases the cooling output, the high-frequency computing task may have already ended. The delayed release of cooling capacity leads to overcooling and energy waste within the micro-module. In the initial stage of a sudden load increase, the cooling system maintains low-power operation because it does not sense the temperature change, causing heat to accumulate rapidly in a localized area. The control mode that uses a large lag process variable to follow high-frequency load changes introduces phase lag mathematically, making it difficult to balance control speed and stability under conditions of drastic load fluctuations.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a control mechanism that breaks through the limitations of physical heat transfer lag, realizes real-time synchronization between control commands and load energy changes, and has the ability to identify load attributes and adapt to environmental parameters. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An industrial control system for a micro-module environmental system, operating at an edge controller, includes a status acquisition module, a logic operation module, and an execution output module: The status acquisition module reads the active power signal of the micro-module power distribution unit in real time and calculates the time change rate of the active power signal. The logic operation module includes a load attribute identification unit and a feedforward timing shaping unit; The load attribute identification unit performs sliding window statistical operations on the time series of the active power signal, calculates the fluctuation variance value of the signal within the current window, and executes a logic gating strategy: when the time change rate of the active power signal exceeds the preset dead zone threshold and the fluctuation variance value is higher than the preset load characteristic threshold, a trigger signal is sent to the feedforward timing shaping unit. The feedforward timing shaping unit, in response to the trigger signal, calculates the theoretical cooling demand value based on the active power signal according to the preset energy mapping rules from electrical energy input to cooling output, and performs smoothing processing on the theoretical cooling demand value using the first-order inertial transfer function to generate a feedforward control command that lags behind the change in active power signal on the time axis; the execution output module receives the feedforward control command and adjusts the operating parameters of the refrigeration equipment in the micro-module.

[0006] Preferably, the logic operation module further includes a structural integrity inference module. The structural integrity inference module calculates in real time the residual sequence between the measured ambient temperature within the micromodule and the theoretically predicted temperature output by the feedforward timing shaping unit. The structural integrity inference module executes residual feature identification logic. When the residual sequence exhibits a persistent steady-state deviation characteristic unrelated to the active power signal, it determines that the micromodule has a physical structural abnormality and triggers the main controller to block the instruction output of the feedforward timing shaping unit and switch to a single feedback control mode.

[0007] Preferably, the execution output module integrates mechanical envelope clamping logic; the mechanical envelope clamping logic stores a safe operating pressure and speed mapping table of the compressor, monitors the condensing pressure and evaporating pressure of the refrigeration equipment in real time, and calculates the maximum allowable safe rate of change under the current pressure state; when the adjustment rate required by the feedforward control command exceeds the maximum allowable safe rate of change, the mechanical envelope clamping logic performs truncation processing on the feedforward control command and outputs a restricted control signal that conforms to the safe operating envelope.

[0008] Preferably, the logic operation module further includes a model parameter adaptive correction unit; the model parameter adaptive correction unit stores a micro-module heat exchange efficiency benchmark model, and calculates the current actual heat exchange efficiency index when the active power signal and the ambient temperature inside the micro-module are both in quasi-steady-state operation; the model parameter adaptive correction unit calculates the deviation between the actual heat exchange efficiency index and the micro-module heat exchange efficiency benchmark model to generate an aging attenuation coefficient, and adjusts the gain coefficient in the energy mapping rule according to the aging attenuation coefficient.

[0009] Preferably, the feedforward timing shaping unit uses a discretized first-order inertial difference equation to smooth the theoretical cooling demand value. The first-order inertial difference equation is as follows: ,in, The feedforward control command output at the current moment. This is the theoretical cooling demand value calculated at the current moment. This refers to the feedforward control command output at the previous moment. It is a smoothing coefficient determined based on the physical thermal time constant of IT equipment, and its value ranges from zero to one.

[0010] Preferably, the system further includes a steady-state correction module; the steady-state correction module acquires the ambient temperature signal within the micro-module and is in an active state within a preset low-frequency sampling period; the steady-state correction module calculates the deviation between the ambient temperature signal and the target temperature, generates an offset correction value based on the PID control algorithm, and superimposes the offset correction value onto the feedforward control command.

[0011] Preferably, the load attribute identification unit stores a load feature fingerprint database; the load feature fingerprint database contains high-frequency random fluctuation feature data of the calculated load and low-frequency smooth fluctuation feature data of the electrical load; the load attribute identification unit compares the calculated fluctuation variance value with the load feature fingerprint database, and when the fluctuation variance value is lower than the preset load feature threshold, it determines that the current power change is a non-thermal load fluctuation, and locks the output of the feedforward control command to maintain the operating state of the refrigeration equipment unchanged.

[0012] Preferably, the mechanical envelope clamping logic also executes a thermal capacity buffer compensation strategy; during the truncation process of the feedforward control command, the thermal capacity buffer compensation strategy monitors the rate of change of ambient temperature within the micro-module; when the rate of change of ambient temperature does not exceed the preset thermal safety limit, the thermal capacity buffer compensation strategy allows the ambient temperature to undergo instantaneous overshoot in order to maintain the refrigeration equipment operating within the safe operating envelope.

[0013] Preferably, the model parameter adaptive correction unit also performs long-term trend analysis on the aging degradation coefficient; when the aging degradation coefficient exceeds the preset maintenance threshold, the model parameter adaptive correction unit generates a maintenance prompt signal indicating a decline in the physical environment performance of the micro-module, and the maintenance prompt signal indicates the physical state of air filter blockage or cold aisle seal failure.

[0014] Preferably, the status acquisition module executes signal denoising logic; the signal denoising logic monitors the time change rate of the active power signal in real time. When the absolute value of the time change rate is greater than five percent per second of the rated power of the micro-module, it determines that an effective load disturbance has occurred and sends a valid data frame to the logic operation module. The system interacts with the power distribution unit and refrigeration equipment through the industrial fieldbus protocol.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the micro-module dynamic environment system, a direct mapping channel from electrical energy input to cooling output is established, and an inertial shaping link simulating physical heat dissipation characteristics is connected in series. At the control logic layer, the timing mismatch of the electrothermal energy conversion process is eliminated, and the traditional temperature feedback lag regulation is transformed into a synchronous response based on energy conservation feedforward. At the instant of a step change in load power, a cooling command overlapping with the time axis of the physical device's heat dissipation curve is generated, avoiding local heat accumulation caused by sensor heat transfer delay, and realizing dynamic synchronization between control commands and physical energy flow.

[0016] 2. The residual sequence is constructed by using the theoretically predicted temperature output of the feedforward model and the actual ambient temperature collected by the sensor. Based on the time-domain characteristics of the residual sequence, physical structure anomalies are diagnosed. The surplus of the internal calculation of the control algorithm is reused to transform the steady-state deviation into a characteristic signal representing the health status of the physical environment. When the residual shows an uncorrelated and persistent steady-state deviation, the system automatically identifies the cold aisle door opening or the airflow short circuit, which damages the integrity of the physical structure and triggers the feedforward logic fuse protection to prevent the control system from blindly increasing the output under physical boundary failure. The existing temperature detection data is used to realize process monitoring and fault self-diagnosis.

[0017] 3. Statistical analysis of the power signal sliding window is performed to extract the second-order moment features of load fluctuations, identify the attributes of computational loads and electrical loads, and utilize the time-domain texture differences between high-frequency random fluctuations of computational tasks and low-frequency smooth fluctuations of battery charging electrical behavior to construct a logic shield for non-thermal load interference. The feedforward cooling response is triggered only when thermal load characteristics are identified, filtering out false power signals caused by grid fluctuations and equipment cold starts, so that the cooling resource input accurately corresponds to the actual calculated heat consumption and avoids ineffective mechanical adjustment. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the feedforward control logic and signal flow of the system of the present invention; Figure 2This is a comparison chart of the time-domain tracking characteristics of the feedforward command and the PID response of the present invention; Figure 3 This is a diagram showing the functional module division and logical architecture of the system of the present invention. Detailed Implementation

[0019] The following examples are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0020] An industrial control system for a micro-modular dynamic environment system includes a state acquisition module, a logic operation module, an execution output module, a steady-state correction module, and a structural integrity inference module. These modules work collaboratively to achieve dynamic synchronization between cooling supply and actual heat load. The state acquisition module constructs the system's feedforward input channel. Given that field signals may contain electromagnetic noise, this module does not directly respond to instantaneous power values ​​but instead reads the active power signal from the power distribution unit in real time. And calculate the time rate of change of the active power signal. The system stores a preset dead-time threshold, for example, set to five percent of the rated power of the micro-module per second. The module executes signal denoising logic. When the absolute value of the calculated time change rate exceeds the preset dead-time threshold, it determines that an effective load disturbance has occurred and sends an effective data frame to the logic operation module. This procedure is used to filter background noise and small non-thermal fluctuations.

[0021] The logic operation module includes a load attribute identification unit and a feedforward timing shaping unit. Addressing the physical differences in time-domain waveforms between computational loads and electrical loads, the load attribute identification unit stores a load feature fingerprint database, containing high-frequency random fluctuation characteristic data of the computational load and low-frequency smooth fluctuation characteristic data of the electrical load. This unit performs sliding window statistical operations on the time series of the active power signal, calculating the fluctuation variance value of the signal within the current window. This unit will calculate the fluctuation variance value. The signal is compared with a preset load characteristic threshold, and a logic gating strategy is executed: when the time rate of change of the active power signal exceeds a preset dead zone threshold, and the fluctuation variance value... When the load characteristic threshold is higher than the preset threshold, a trigger signal is sent to the feedforward timing shaping unit. If the fluctuation variance value... If the load is below a preset load characteristic threshold, the unit determines that the current power change is a non-thermal load fluctuation such as battery float charging, and locks the output of the feedforward control command to maintain the operating state of the refrigeration equipment. The feedforward timing shaping unit is used to solve the timing mismatch problem between energy input and heat release. This unit responds to the trigger signal and, according to a preset energy mapping rule from electrical energy input to cooling output, such as a linear function, shapes the load. Based on active power signal Calculate the theoretical cooling demand value To simulate the physical inertia of heat dissipation, this unit uses a discretized first-order inertial difference equation to smooth the theoretical cooling demand value. The equation is as follows: ,in The feedforward control command output at the current moment. This is the theoretical cooling demand value calculated at the current moment. This refers to the feedforward control command output at the previous moment. The smoothing coefficient is determined based on the physical thermal time constant of the IT equipment, and its value ranges from zero to one. The instructions generated by this calculation lag behind the power change on the time axis and remain synchronized with the equipment's heat release curve.

[0022] The output module receives feedforward control commands and adjusts the refrigeration equipment. To prevent commands from exceeding the physical limits of the actuator, this module integrates mechanical envelope clamping logic, stores a safe operating pressure-speed mapping table for the compressor, and monitors the condensing pressure of the refrigeration equipment in real time. With evaporation pressure And calculate the maximum permissible safe rate of change under the current pressure state. When the adjustment rate required by the feedforward control command exceeds the maximum permissible safe rate of change, this logic truncates the feedforward control command and outputs a restricted control signal that conforms to the safe operating envelope. During this period, a thermal capacity buffer compensation strategy is also implemented to monitor the rate of change of ambient temperature within the micromodule. When this rate of change does not exceed the preset thermal safety limit, a momentary overshoot of the ambient temperature is allowed to maintain the refrigeration equipment operating within the safe operating envelope. Regarding the preset thermal safety limit involved in the thermal capacity buffer compensation strategy, this limit is not a fixed value, but is determined by the system based on the real-time monitored dew point temperature. The rate of temperature rise allowed in the specifications of IT equipment manufacturers Dynamic calculations show that the system defines the difference between the current return air temperature and the dew point temperature as the safety margin, and calculates the maximum allowable instantaneous temperature rise under the premise that condensation does not occur and the equipment overheat protection is not triggered, based on the equipment's heat capacity characteristics. ,Should This constitutes a hard constraint boundary when the heat capacity buffer compensation strategy is executed.

[0023] The logic operation module also includes a structural integrity inference module for diagnosing physical environment anomalies. This module calculates the measured ambient temperature within the micromodule in real time. The theoretical predicted temperature output of the feedforward timing shaping unit residual sequences between This module executes residual feature identification logic, when the residual sequence When a persistent steady-state deviation characteristic unrelated to the active power signal is observed, it is determined that the micromodule has a physical structural abnormality such as an open cold aisle door or airflow short circuit. At this time, the main controller is triggered to block the command output of the feedforward timing shaping unit and switch to a single feedback control mode. The logic operation module also includes a model parameter adaptive correction unit for compensating for physical aging. This unit stores a benchmark model of the micromodule's heat exchange efficiency. When the active power signal and the ambient temperature inside the micromodule are both detected to be in quasi-steady-state operation, the current actual heat exchange efficiency index is calculated. This unit calculates the deviation between the actual index and the benchmark model to generate an aging degradation coefficient. And adjust the gain coefficient in the energy mapping rule according to this coefficient. When the aging degradation coefficient When the preset maintenance threshold is exceeded, the unit generates a physical condition maintenance prompt signal indicating air filter blockage or cold aisle seal failure. The system also includes a steady-state correction module to eliminate steady-state errors. This module collects the ambient temperature signal within the micromodule and remains active within a preset low-frequency sampling period. The module calculates the deviation between the ambient temperature signal and the target temperature and generates an offset correction value based on a PID control algorithm. The bias correction value is superimposed on the feedforward control command, and the final output command is obtained. .

[0024] Example 1: In a high-frequency quantitative trading center scenario, the high-density server racks within the micro-module need to handle the computational load surge from idle to full load within minutes. Under this condition, the instantaneous power consumption of IT equipment... Within 10 seconds, the power surged from the baseline of 2kW to 18kW. Traditional temperature sensors, due to the thermal inertia of the air medium, typically only detect the rise in return air temperature 60 to 90 seconds after the load surge. This physical time difference causes the refrigeration system to be in a low-power standby state in the early stage of heat accumulation, and then maintain a high cooling output due to the lagging temperature signal after the heat load subsides, resulting in overshoot of cooling capacity and energy waste.

[0025] When the system faces the above operating conditions, the status acquisition module samples the active power signal of the power distribution unit at a frequency of 5Hz. Once calculated If the power fluctuation exceeds the preset dead zone threshold of 0.5 kW / s, the logic operation module will immediately start. At this time, the load attribute identification unit will determine the power fluctuation variance within the current time window. Perform the calculation, if If the power surge exceeds a preset load characteristic threshold, the system determines that the surge originates from a computing task rather than electrical noise, and then triggers a feedforward response. The feedforward timing shaping unit then... Calculate the theoretical cooling demand and apply a smoothing coefficient. The first-order inertial difference equation is set to 0.05. The system generates control commands to drive the refrigeration compressor to increase its speed according to a preset exponential curve before the temperature sensor detects a temperature rise. During this process, the mechanical envelope clamping logic monitors the condensing pressure in real time. With evaporation pressure When the calculated maximum permissible safety change rate When limiting the compressor's acceleration slope, the system automatically invokes a thermal capacity buffer compensation strategy, allowing for a short-term, minute overshoot of the ambient temperature within the thermal safety limits to ensure the absolute safety of mechanical components. As the computational task concludes, the power signal... steep descent, feedforward command The cooling capacity then drops rapidly, avoiding the long-term redundancy of cooling capacity caused by the integral action of traditional PID control. This feedforward synchronization mechanism based on the energy input end eliminates physical heat transfer lag, realizes dynamic matching between cooling supply and actual heat load, and avoids the dual risks of local hot spots and energy waste.

[0026] Example 2: This example aims to verify the dynamic response characteristics and engineering stability of the aforementioned industrial control system under complex operating conditions. The experiment was conducted on a full physical simulation platform for a micro-module data center with a rated power of 200kW, equipped with a programmable electronic load group to simulate the power consumption changes of IT equipment, and a variable frequency row-level precision air conditioning system. Data acquisition was performed using a power analyzer with an accuracy of 0.2 to read the electrical parameters of the power distribution circuit, and a T-type thermocouple grid with a response time of less than 500ms was used to monitor the temperature field distribution of the hot and cold aisles. The sampling period for the core control parameters was set to 200ms, and this parameter was determined based on Shannon's sampling theorem and... The engineering trade-off for the response bandwidth of the cold actuator aims to ensure the ability to capture millisecond-level transient characteristics of IT loads while avoiding inverter oscillations caused by high-frequency noise introduced by oversampling. An experimental control group and the prototype of this invention were established. The control group employed a traditional control strategy based on PID feedback of return air temperature, while the prototype of this invention utilized a complete set of feedforward control logic, including load attribute identification, timing shaping, and mechanical envelope clamping functions. The experimental process included two consecutive dynamic test phases, designed to comprehensively evaluate the system's ability to identify and respond to effective and spurious heat loads. In the first phase, the electronic load group executed a step load command, with active power... The power output linearly increases from 20kW to 180kW within 5.0 seconds to simulate sudden computational tasks in a high-frequency trading scenario. To verify the system's reliability under strong electromagnetic interference, Gaussian white noise with a signal-to-noise ratio of 20dB is superimposed on the power signal source. In the second stage, the electronic load group maintains a constant power, but a low-frequency smoothing power fluctuation with a frequency of 0.1Hz and an amplitude of 15kW is introduced by simulating a battery float charging circuit. During this process, the status acquisition module continuously monitors... The logic operation module calculates the fluctuation variance value in real time. And compare it with the preset load characteristic threshold.

[0027] The recording and analysis of experimental data revealed the differences between the two control strategies at the physical mechanism level. In the first stage, when the load increased dramatically, the control group, due to the thermal hysteresis of the air medium, maintained a low speed for 45 seconds after the load surge, causing the hot aisle temperature to rapidly accumulate and exceed the safety limit. The PID controller, due to accumulated errors, output an oversaturation command, resulting in severe temperature overshoot. In contrast, the sample group of this invention detected… Exceeding the dead zone threshold and After exhibiting high-frequency random texture characteristics, the effective heat load was immediately confirmed. The feedforward timing shaping unit then output a speed command that increased exponentially. This command curve matched the heat release curve of the heat sink measured by the thermocouple on the time axis. It is worth noting that when the speed change rate required by the command approached the compressor surge boundary, the mechanical envelope clamping logic limited the acceleration slope. Although this caused a slight increase in instantaneous temperature, it avoided drastic fluctuations in the compressor discharge pressure. In the second stage of the sham heat load test, although the power reading increased, due to... When the value is below the threshold, the sample group of the present invention correctly locks the feedforward output to avoid ineffective cooling; while the control group, due to the lack of attribute identification ability, still performs unnecessary adjustment actions after the temperature fluctuates slightly due to electrical heating. Table 1 shows the comparison data of key performance indicators under the two control modes.

[0028] Table 1: Comparison of Control Performance under Step Load and Dummy Load Conditions ; Referring to Table 1, the data shows that the prototype of this invention shortens the cooling response delay by two orders of magnitude, reduces the maximum temperature deviation by 78.5%, and controls the compressor discharge pressure fluctuation rate within a safe range. Under non-thermal load interference, the prototype of this invention achieves zero malfunctions. Through energy flux mapping and virtual heat capacity shaping mechanism, the system can effectively avoid physical heat conduction lag and achieve precise synchronization between cooling supply and actual heat load.

[0029] Example 3: This example combines Figures 1 to 3 A description of an industrial control system for a micro-module environmental system, such as... Figure 1As shown, the system's data flow originates from the active power signal source output by the power distribution unit. This signal is transmitted to the status acquisition module to perform signal denoising logic and rate of change calculation to generate a valid data frame. The data enters the load attribute identification unit to perform sliding window variance statistics and feature comparison to generate a trigger signal. The feedforward timing shaping unit receives the trigger signal and generates a feedforward control command based on energy mapping and first-order inertial smoothing. This command is transmitted to the execution output module and processed by mechanical envelope clamping logic and thermal capacity buffering strategy to be converted into a restricted control signal. Ultimately, this signal drives the refrigeration equipment to adjust the compressor or fan parameters. At the same time, the model parameter adaptive correction unit calculates the aging attenuation coefficient and performs gain correction to achieve parameter correction. Meanwhile, the structural integrity inference module performs residual feature identification and physical anomaly diagnosis to trigger abnormal fuses when necessary.

[0030] like Figure 2 As shown in the chart, this graph illustrates the dynamic relationship of three key variables on the time axis. The horizontal axis represents time in seconds, and the vertical axis represents numerical values. The solid line depicts the trajectory of active power P(t) in kW as the load changes stepwise, and the dashed line represents the feedforward control command. The response curve closely follows power changes but exhibits smooth characteristics, while the dotted line reflects the time lag of the traditional PID response. A comparison shows that the feedforward control command can track power changes more quickly on both the rising and falling edges, and maintains a steady-state output synchronized with the power curve at the peak. Figure 3 As shown, the system's functional architecture is logically divided into a status acquisition module, an execution output module, a load attribute identification unit, and a feedforward shaping and correction section. The status acquisition module is responsible for acquiring active power signals and time change rate. The calculation and signal denoising logic, the output module integrates safety rate of change truncation, thermal capacity buffer compensation strategy and mechanical envelope clamping logic, and the load attribute identification unit performs sliding window statistical calculations and fluctuation variance values. The calculation and load feature fingerprint database comparison, the core feedforward shaping and correction part covers structural integrity inference, adaptive correction of model parameters, energy mapping rules and first-order inertial transfer function. All modules work together to ultimately point to a single control goal, that is, to achieve precise dynamic synchronization of cooling supply and heat load.

[0031] Example 4: This example addresses the problem of energy mapping rule failure caused by physical environment aging during long-term system operation, and the resulting black-box problem of setting and adjusting the aging attenuation coefficient—a key parameter in the adaptive correction unit of model parameters. It constructs an online parameter identification and adaptive correction procedure based on quasi-steady-state data analysis to solve the engineering problem of fixed control models being unable to adapt to time-varying physical environments, ensuring control accuracy throughout the system's lifecycle. In the long-term operation of the micro-module dynamic environment system, physical degradation phenomena such as dust accumulation on air filters, fouling of heat exchangers, and aging of cold aisle seals are inevitable. These physical changes directly affect the overall thermal resistance of the micro-module. As time increases, if the control system continues to use the factory-preset fixed energy mapping rules... Under the same IT load power Under these conditions, the cooling capacity output by the system will be insufficient to maintain the expected thermal balance, leading to steady-state temperature drift. This causes the PID feedback loop to remain in a high-load correction state for an extended period, compressing the system's dynamic adjustment margin. To quantify this physical degradation process and dynamically compensate the control model, this embodiment introduces an aging degradation coefficient. As the core correction factor, the determination and application process is as follows: the system executes quasi-steady-state identification logic to capture an effective parameter calibration window, and the state acquisition module continuously monitors the active power signal. With ambient temperature signal The real-time variance, when within a preset time window length For example, within 15 minutes, the power signal variance Less than the steady-state power threshold And temperature signal variance Less than the steady-state temperature threshold When the system determines that it is currently in a quasi-steady-state operating mode, in this mode, the system calculates the current actual heat exchange efficiency index. , defined as the sensible heat load that can be removed per unit cooling power, is calculated using the following formula: ,in This represents the current average power of the IT load. The system uses the average input power of the current refrigeration equipment and simultaneously retrieves a preset micro-module heat exchange efficiency benchmark model. The efficiency curves were obtained by traversing different load rates under standard factory operating conditions.

[0032] The system calculates the relative deviation between the actual performance indicators and the benchmark model output values ​​to generate the aging degradation coefficient. The specific calculation formula is as follows: The coefficient Directly characterizes the degree of performance degradation of the current physical environment relative to its initial state; for example, if A calculation result of 0.1 means that the current heat exchange efficiency is 10% lower than the factory standard. To avoid parameter fluctuations caused by measurement noise, the system adjusts the calculated value. The values ​​are processed using an exponentially weighted moving average (EWMA) filter to generate smoothed correction coefficients. Finally, the model parameter adaptive correction unit is based on the smoothed aging decay coefficient. The energy mapping rule in the dynamically reconstructed feedforward control law, and the corrected gain coefficient. Updated to This ensures that when the thermal resistance of the physical environment increases, leading to a decrease in efficiency, the control system can automatically increase the feedforward gain and output a larger theoretical cooling demand value. This allows the performance degradation at the logical level to be offset by the performance degradation at the physical level. In addition, the system has a maintenance warning threshold set. For example, 0.25, when If the gain coefficient continues to exceed this threshold, the system will no longer increase the gain coefficient, but will instead trigger a maintenance prompt signal, instructing maintenance personnel to check the air filter or sealing components to prevent the system from operating with serious physical faults.

[0033] Example 5: During the initial deployment phase of the system, to ensure the accurate matching of model parameters with the current physical environment, a set of on-site pre-deployment calibration procedures are executed. The chiller unit is started under no-load conditions and a series of stepped reference speeds are set. At the same time, the corresponding steady-state input power is recorded using a high-precision power analyzer. During this process, a known constant heat source is simulated using a standard heat load box, and the steady-state response value of the ambient temperature is collected. This constructs an initial micro-module heat exchange efficiency benchmark model, establishes the mapping relationship between different cooling powers and heat removal capabilities under ideal operating conditions, and provides an absolute reference benchmark for subsequent online adaptive correction. This step eliminates the initial model deviation caused by equipment manufacturing tolerances or installation differences.

[0034] The aging degradation coefficient involved in the adaptive correction unit for model parameters To determine the appropriate parameters, an offline calibration and data filling procedure was established. This involved artificially simulating different levels of physical aging scenarios in a controlled laboratory environment. For example, using standard shielding nets to simulate different levels of filter clogging or adjusting the thickness of the thermal resistance coating on the heat exchanger surface. This yielded a series of corresponding actual heat exchange performance indicators. By comparing these measured indicators with a benchmark model, the aging degradation coefficient was systematically calibrated. The quantitative functional relationship between physical aging and the degree of physical degradation is established and solidified into a core data table stored in the controller's non-volatile memory. This data table forms the mathematical basis of the online correction logic, ensuring that the system can index the correct correction coefficient based on real-time monitoring data when facing unknown levels of physical degradation.

[0035] Example 6: During the system's on-site deployment and initialization calibration phase, the smoothing coefficient of the feedforward timing shaping unit is addressed. Perform a thermal response step test: Control the micro-module IT load from no load to 50% of rated power in a step load, record the ambient temperature response curve using a return air temperature sensor, and define the time required for the temperature rise to reach 63.2% of the steady-state value as the physical thermal time constant. According to the formula calculate, The sampling period of the edge controller system is used, and the calculation results are solidified into feedforward filter parameters. At the same time, load characteristic thresholds are preset for the load attribute identification unit. During non-computation task periods, the active power sequence of the battery float charging and standby states of the power distribution unit is collected. The variance of the sequence sliding window fluctuation is calculated and the upper limit of the 99% confidence interval is taken as the benchmark threshold. The data is written into non-volatile memory to quantify and define the statistical boundary of non-thermal load fluctuation.

[0036] The mechanical envelope clamping logic compressor's safe operating pressure and speed mapping table is constructed based on the refrigeration compressor manufacturer's standard operating envelope curve. It is discretized to form a two-dimensional lookup table stored in the controller. The two-dimensional lookup table uses condensing pressure... and evaporation pressure As an indexed input, the output should contain the maximum permissible speed and the speed change rate limit for the corresponding operating condition. During system operation, the output module reads data at 100ms intervals. and The real-time value is obtained by looking up the table to retrieve the current physical boundary constraints. When the limit value is less than the feedforward command value, the feedforward command is hard-limited to prioritize maintaining the compressor in the safe intake superheat and exhaust temperature range.

[0037] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An industrial control system for a micro-module moving ring system, running on an edge controller, characterized in that, The system comprises a state acquisition module, a logic operation module, and an execution output module. The state acquisition module reads the active power signal of the micro-module power distribution unit in real time, and calculates the time rate of change of the active power signal. The logic operation module comprises a load attribute identification unit and a feedforward timing shaping unit. The load attribute identification unit performs a sliding window statistical operation on the time series of the active power signal, calculates the fluctuation variance value of the signal in the current window, and performs a logic gating strategy: when the time rate of change of the active power signal exceeds a preset dead zone threshold and the fluctuation variance value is higher than a preset load characteristic threshold, a trigger signal is sent to the feedforward timing shaping unit. The feedforward timing shaping unit, in response to the trigger signal, calculates a theoretical refrigeration demand value based on the active power signal according to a preset energy mapping rule from electric energy input to cold energy output, and performs smoothing processing on the theoretical refrigeration demand value using a first-order inertia transfer function to generate a feedforward control instruction that lags behind the change of the active power signal in the time axis; the execution output module receives the feedforward control instruction and adjusts the operating parameters of the refrigeration equipment in the micro-module.

2. The industrial control system of claim 1, wherein, The logic operation module further comprises a structural integrity inference module that calculates the residual sequence between the measured ambient temperature in the micro-module and the theoretically predicted temperature output by the feedforward timing shaping unit in real time; the structural integrity inference module performs residual feature identification logic, and when the residual sequence presents a persistent steady-state deviation feature that is unrelated to the active power signal, it is determined that there is a physical structure abnormality in the micro-module, and the output of the feedforward timing shaping unit is blocked by the main controller, switching to a single feedback control mode.

3. The industrial control system of claim 1, wherein, The execution output module integrates a mechanical envelope clamping logic; the mechanical envelope clamping logic stores a compressor safe operating pressure and speed mapping table, monitors the condensing pressure and evaporating pressure of the refrigeration equipment in real time, and calculates the maximum allowable safe change rate under the current pressure state; when the adjustment rate required by the feedforward control instruction exceeds the maximum allowable safe change rate, the mechanical envelope clamping logic performs truncation processing on the feedforward control instruction, and outputs a limited control signal that meets the safe operating envelope.

4. The industrial control system of claim 1, wherein, The logic operation module further comprises a model parameter adaptive correction unit; the model parameter adaptive correction unit stores a micro-module heat exchange efficiency benchmark model, and calculates the actual heat exchange efficiency index when it is detected that both the active power signal and the ambient temperature in the micro-module are in quasi-steady-state operation; the model parameter adaptive correction unit calculates the deviation between the actual heat exchange efficiency index and the micro-module heat exchange efficiency benchmark model to generate an aging attenuation coefficient, and adjusts the gain coefficient in the energy mapping rule according to the aging attenuation coefficient.

5. The industrial control system of claim 1, wherein, The feedforward timing shaping unit performs smoothing processing on the theoretical refrigeration demand value using a discretized first-order inertial difference equation, which is: wherein, is the feedforward control instruction output at the current time, is the theoretical refrigeration demand value calculated at the current time, is the feedforward control instruction output at the previous time, is a smoothing coefficient determined based on the physical thermal time constant of the IT equipment, and has a value ranging from zero to one.

6. The industrial control system of claim 1, wherein, The system further comprises a steady-state correction module; the steady-state correction module collects the ambient temperature signal in the micro-module and is in an activated state in a preset low-frequency sampling period; the steady-state correction module calculates the deviation of the ambient temperature signal from the target temperature, and generates a bias correction value based on a PID control algorithm, and superimposes the bias correction value on the feedforward control instruction.

7. The industrial control system of claim 1, wherein, The load attribute identification unit stores a load characteristic fingerprint library. The load characteristic fingerprint library contains high-frequency random fluctuation characteristic data of the calculated load and low-frequency smooth fluctuation characteristic data of the electrical load; the load attribute identification unit compares the calculated fluctuation variance value with the load characteristic fingerprint library, and when the fluctuation variance value is lower than the preset load characteristic threshold, it is determined that the current power change is a non-thermal load fluctuation, and the output of the feedforward control instruction is locked to maintain the operating state of the refrigeration equipment unchanged.

8. The industrial control system of claim 3, wherein, The mechanical envelope clamping logic also performs a thermal capacity buffer compensation strategy; the thermal capacity buffer compensation strategy monitors the rate of change of the ambient temperature in the micro-module during the execution of the feedforward control instruction truncation process; when the rate of change of the ambient temperature does not exceed the preset thermal safety limit, the thermal capacity buffer compensation strategy allows the ambient temperature to have a transient overshoot to maintain the refrigeration equipment operating within the safety operating envelope.

9. The industrial control system of claim 4, wherein, The model parameter adaptive correction unit also performs long-term trend analysis on the aging attenuation coefficient; when the aging attenuation coefficient exceeds the preset maintenance threshold, the model parameter adaptive correction unit generates a maintenance prompt signal indicating the decline in the physical environment performance of the micro-module, and the maintenance prompt signal indicates the physical state of the air filter blockage or the failure of the cold channel seal.

10. The industrial control system of claim 1, wherein, The state acquisition module performs signal denoising logic; the signal denoising logic monitors the time rate of change of the active power signal in real time, and when the absolute value of the time rate of change is greater than five percent of the rated power of the micro-module per second, it is determined that an effective load disturbance has occurred and an effective data frame is sent to the logic operation module, and the system interacts with the power distribution unit and the refrigeration equipment through the industrial field bus protocol.

Images