A dynamic temperature control system for a refrigerated centrifuge

Abstract

The application provides a dynamic control temperature system of a refrigeration centrifuge, which is composed of a multi-dimensional heat state sensing module, a thermodynamic modeling and estimation module, a temperature control decision module, a refrigeration output execution module, a temperature control record generation module and an abnormal information distribution module; the application realizes multi-dimensional heat state sensing fusion, collects rotor speed, cavity multi-point temperature, equipment operation parameters and environmental parameters and other multi-dimensional heat characteristic data, makes up for the data limitations of traditional single-point temperature measurement, provides more comprehensive data support for equipment heat balance deduction and temperature trend estimation, improves the accuracy of temperature control prediction, and is suitable for adaptive regulation and control of the centrifuge in the whole running stage, can match the corresponding control model according to the heat load characteristics of different working conditions such as speed-up, constant speed and speed-down, carries out pre-cooling compensation in the speed-up stage, switches to an energy-saving maintenance mode in the constant speed stage, can relieve the response lag problem of traditional feedback control, and is suitable for the temperature control requirements of different working conditions.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for refrigerated centrifuges, and in particular to a dynamic temperature control system for refrigerated centrifuges. Background Technology

[0002] Refrigerated centrifuges are core process equipment in biopharmaceutical, medical testing, and chemical separation fields. They primarily rely on the centrifugal force field generated by the high-speed rotation of the rotor to separate material components. To ensure the activity of biological samples, avoid changes in the physicochemical properties of materials, and improve the purity of separated products, the equipment must rely on a refrigeration system to stably maintain the centrifuge chamber within a preset low-temperature range during operation. The operating accuracy and response performance of the temperature control system have a critical impact on the quality of the finished product and the stability of the process. Currently, most temperature control solutions for refrigerated centrifuges in the industry adopt a control method that combines single-point or a few-point feedback from temperature sensors with PID regulation and hysteresis comparison. By detecting the difference between the actual temperature of the chamber and the target temperature, the compressor operating power or the opening of the electronic expansion valve is adjusted accordingly. To optimize the temperature control effect, many improved solutions have gradually emerged. Some solutions can predict equipment heat load fluctuations based on changes in ambient temperature and slightly adjust the PID control parameters. Other solutions attempt to improve chamber temperature fluctuations under low-load conditions by adding auxiliary heating structures to the evaporator pipeline. Overall, these solutions have achieved certain optimization effects on basic temperature control stability.

[0003] While existing improved temperature control solutions can generally meet basic temperature control requirements under normal steady-state conditions, they still suffer from insufficient adaptability and intelligence in complex, variable-condition centrifugal operations. This is primarily reflected in four dimensions: thermal state perception, condition-adaptive control, process scheduling, and anomaly handling. First, existing temperature control perception systems have a relatively limited scope, mostly relying solely on chamber temperature data for regulation, with less consideration given to dynamic changes in rotor speed, heat generation from air friction, differences in material heat capacity, and fluctuations in ambient temperature and humidity. This may result in incomplete reference data for temperature trend prediction. Second, existing control parameters are mostly fixed setpoints or only slightly adjusted to adapt to ambient temperature, making it difficult to match the differentiated heat load characteristics at different operating stages of the centrifuge (acceleration, constant speed, and deceleration). During rotor acceleration, the rate of heat generation from air resistance is rapid, and traditional feedback control modes may exhibit some response lag, easily leading to small temperature overshoot. Furthermore, the temperature control logic is relatively independent of the centrifugal process flow, making it difficult to proactively predict heat load changes based on preset process parameters, resulting in a degree of passivity in refrigeration resource allocation. In addition, the existing equipment has a relatively simple temperature control anomaly handling mode, which relies heavily on manual alarm intervention. The ability to automatically identify, degrade the handling and trace data needs to be improved. Therefore, this invention proposes a dynamic temperature control system for refrigerated centrifuges. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic temperature control system for a refrigerated centrifuge to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a dynamic temperature control system for a refrigerated centrifuge, comprising a multi-dimensional thermal state sensing module, a thermodynamic modeling and prediction module, a temperature control decision module, a refrigeration output execution module, a temperature control record generation module, and an anomaly information distribution module; wherein: The multidimensional thermal state sensing module is equipped with a comprehensive sensor data acquisition unit. After completing the acquisition of the operating status and thermal parameters, the unit actively broadcasts an acquisition ready signal. The acquisition ready signal is sent to the event listening unit inside the temperature control decision module. After receiving the signal, the event listening unit immediately triggers the process coordination unit configured in the temperature control decision module to enter the working state. The thermodynamic modeling and prediction module is equipped with a thermal balance deduction unit. The feature analysis and matching unit deployed in the temperature control decision module obtains output data from the integrated sensor data acquisition unit, and obtains multi-dimensional thermal feature indicators by solving the output data. Then, the feature analysis and matching unit converts the solved multi-dimensional thermal feature indicators into thermal balance deduction input vectors according to the preset thermal feature type and deduction model comparison table, and sends the input vectors to the thermal balance deduction unit. After the process coordination unit is started, it executes two parallel tasks simultaneously: first, it calls the heat balance simulation unit to perform heat load estimation and cooling capacity demand calculation; second, it calls the cooling output execution module to perform dynamic cooling capacity pre-allocation. After the heat load estimation, cooling capacity demand calculation and dynamic cooling capacity pre-allocation are completed, the process coordination unit triggers the temperature control record generation module to perform the operation record collection operation. After the temperature control record generation module completes the operation record collection, it links the abnormal information distribution module to perform the abnormal event distribution operation.

[0006] Preferably, the multi-dimensional thermal characteristic indicators collected by the integrated sensor data acquisition unit include at least the current rotor speed, multi-point temperature distribution in the centrifugal cavity, external ambient temperature, real-time operating frequency of the compressor, pressure value of the refrigeration circuit, and the current operating stage indicator; the thermal characteristic type and deduction model comparison table uses the operating stage, speed range, and temperature change gradient as search conditions to configure the corresponding thermal balance deduction model parameter set.

[0007] Preferably, when the current rotor speed is in the acceleration phase and the cavity temperature rise rate exceeds the preset rise rate threshold, the lookup table output includes characteristic parameters such as the thermal load prediction coefficient of the acceleration phase and the pre-cooling compensation amplitude; when the current rotor speed is in the constant speed stable phase and the cavity temperature fluctuation amplitude is less than the preset stability threshold, the lookup table output includes characteristic parameters such as the thermal balance correction coefficient of the constant temperature phase and the energy-saving operation parameters.

[0008] Preferably, the process coordination unit continuously monitors the completion status of the two parallel tasks, and updates the working status to "in execution" after the heat load estimation, cooling demand calculation and cooling pre-allocation are completed, triggering the cooling output execution module to output dynamic cooling instructions; after the cooling allocation is completed, the working status is updated to "pending archiving", and the abnormal information distribution module is linked to perform abnormal event distribution.

[0009] Preferably, the temperature control decision module further includes an operation status tracking unit. The operation status tracking unit uses time points as indexes to record the operation data and dwell time of each sampling moment in the multi-dimensional thermal state perception, thermodynamic modeling, and cooling capacity allocation stages, forming an operation log for backtracking the temperature control decision process.

[0010] Preferably, the temperature control decision module further includes a fault diagnosis and flexible recovery unit, which is used to reallocate the cooling capacity in a graded reduction manner according to a preset flexible recovery strategy when the process coordination unit calls the cooling output execution module and the response timeout, temperature deviation exceeds the threshold, or sensor data is abnormal; if the flexible recovery is successful, the result is returned; if the flexible recovery fails, the abnormality is marked for maintenance and an alarm event is generated and distributed.

[0011] Preferably, the temperature control decision module further includes an experience self-learning unit, and the abnormal information distribution module is provided with an event encapsulation and notification subunit and a feedback collection subunit; the experience self-learning unit receives the manually reviewed and labeled data returned by the feedback collection subunit and associates and matches the original thermal feature parameters, reduces the matching weight of the inference model parameters with large prediction deviations after review and confirmation, and adds the parameter combinations that are confirmed to be effective after manual correction to the lookup table and assigns them initial weights.

[0012] Preferably, the temperature control decision module further includes a local edge computing unit, which is used to deploy the heat balance calculation service, multi-dimensional sensor data parsing service and intermediate data in the scheduling process in the local embedded controller environment of the centrifuge, and isolate data interaction with the external public network.

[0013] Preferably, the multi-dimensional thermal state sensing module is further provided with a sensor data preprocessing and validity verification unit, which is used to filter and smooth the raw data from multiple sensors and remove outliers before handing it over to the integrated sensor data acquisition unit for feature extraction; the operation status tracking unit divides the status of each temperature control cycle into six status nodes for tracking: pending acquisition, feature parsing, deduction calculation, allocation decision, execution confirmation, and archived.

[0014] Preferably, the thermal feature type and deduction model comparison table covers four operating stages: acceleration stage, constant speed stage, deceleration stage, and standby stage; when the feature analysis and matching unit identifies that the temperature rise rate in the acceleration stage exceeds 0.5℃ / second or the temperature fluctuation amplitude in the constant speed stage exceeds ±0.3℃, it activates and outputs the specific deduction model parameter set for the corresponding stage.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention achieves multi-dimensional thermal state perception fusion. By collecting multi-dimensional thermal characteristic data such as rotor speed, multi-point temperature of the cavity, equipment operating parameters and environmental parameters, it makes up for the data limitations of traditional single-point temperature measurement. It can provide more comprehensive data support for equipment thermal balance simulation and temperature trend prediction, and improve the accuracy of temperature control prediction.

[0016] 2. Adaptive control for all stages of centrifuge operation: It can match the corresponding control model according to the heat load characteristics of different operating conditions such as acceleration, constant speed and deceleration. Pre-cooling compensation is carried out during the acceleration stage, and energy-saving maintenance mode is switched during the constant speed stage. This can alleviate the response lag problem of traditional feedback control and adapt to the temperature control requirements of different operating conditions.

[0017] 3. Construct a collaborative scheduling system for the entire temperature control process, connecting all aspects of thermal state perception, load prediction, refrigeration control, and data recording, to achieve linkage and adaptation between temperature control logic and centrifugal process flow, and improve the overall automated closed-loop operation capability of the temperature control process.

[0018] 4. It has the ability to handle abnormal situations flexibly and learn from experience. It can carry out graded and downgraded processing for abnormal temperature control scenarios, reduce the impact of single-point abnormalities on the overall temperature control system, and continuously optimize the control model based on historical operating data to gradually improve the temperature control accuracy and adaptability to complex working conditions of the equipment. Attached Figure Description

[0019] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein: Figure 1 This is a block diagram of the dynamic temperature control system for the refrigerated centrifuge of the present invention; Figure 2 This is a module association diagram of the dynamic temperature control system for the refrigerated centrifuge of the present invention; Figure 3 This is a flow diagram of the temperature control cycle state nodes of the present invention; Numbering on the map: 100. Multidimensional thermal state sensing module; 110. Sensor data preprocessing and validity verification unit; 120. Integrated sensor data acquisition unit; 200. Thermodynamic Modeling and Prediction Module; 210. Thermal Balance Deduction Unit; 300. Temperature control decision-making module; 310. Event monitoring unit; 320. Feature parsing and matching unit; 330. Process coordination unit; 340. Operation status tracking unit; 350. Fault diagnosis and elastic recovery unit; 360. Local edge computing unit; 370. Experience self-learning unit; 400. Cooling output execution module; 500. Temperature control record generation module; 600. Exception information distribution module; 610. Event encapsulation and notification subunit; 620. Feedback collection subunit. Detailed Implementation

[0020] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0021] Example 1 According to one embodiment of the present invention, Figures 1-3 As shown, a dynamic temperature control system for a refrigerated centrifuge includes a multi-dimensional thermal state sensing module 100, a thermodynamic modeling and prediction module 200, a temperature control decision module 300, a refrigeration output execution module 400, a temperature control record generation module 500, and an anomaly information distribution module 600. The temperature control decision module 300 establishes bidirectional communication connections with the multi-dimensional thermal state sensing module 100, the thermodynamic modeling and prediction module 200, the refrigeration output execution module 400, the temperature control record generation module 500, and the anomaly information distribution module 600, respectively, to coordinate the various modules in a unified manner to complete the fully automated temperature control operation, including multi-dimensional sensing acquisition, thermal feature calculation, thermal balance deduction, refrigeration capacity allocation, operation record collection, and anomaly event distribution.

[0022] The temperature control decision module 300 includes an event listening unit 310, a feature parsing and matching unit 320, a process coordination unit 330, an operation status tracking unit 340, a fault diagnosis and elastic recovery unit 350, a local edge computing unit 360, and an experience self-learning unit 370.

[0023] The thermodynamic modeling and prediction module 200 has a heat balance deduction unit 210, which is used to respond to heat balance deduction requests and generate heat load prediction results and cooling capacity demand suggestions.

[0024] The multidimensional thermal state sensing module 100 is equipped with a sensor data preprocessing and validity verification unit 110 and a comprehensive sensor data acquisition unit 120.

[0025] The exception information distribution module 600 has an event encapsulation and notification subunit 610 and a feedback collection subunit 620.

[0026] Internal coordination and operation process of this system: 1. Event Monitoring and Process Triggering: The event monitoring unit 310 establishes an event subscription relationship with the integrated sensor data acquisition unit 120, continuously monitoring the execution status of the sensor acquisition task. After completing the acquisition of multi-dimensional thermal characteristic data for one temperature control cycle, the integrated sensor data acquisition unit 120 sends out a ready signal carrying a timestamp, current rotor speed, multi-point temperature distribution in the centrifuge chamber, external ambient temperature, real-time compressor operating frequency, refrigeration circuit pressure value, and a marker of the current operating stage. Upon receiving this signal, the event monitoring unit 310 forwards the signal to the process coordination unit 330 within a preset response window (e.g., 2 seconds), triggering the downstream deduction and allocation process to start, without requiring any manual triggering action from the operator on the interface. If no ready signal is received within the preset response window and the preset waiting time limit (e.g., 10 seconds) has been exceeded, the event monitoring unit 310 determines that the acquisition task has timed out, generates a timeout warning, and forwards it to the fault diagnosis and flexible recovery unit 350 for handling.

[0027] 2. Thermal Feature Analysis and Matching: The feature analysis and matching unit 320 is used to solve the multi-dimensional sensing data output by the integrated sensing data acquisition unit 120 into the structured input vector required for thermal equilibrium deduction. Its matching rules are as follows: Figure 3 As shown. Specifically, the feature parsing and matching unit 320 extracts the rotor current speed, cavity temperature change gradient, external ambient temperature deviation, compressor current frequency, and operating stage indicators from the sensor data. Based on the preset thermal feature type and deduction model comparison table, it converts the operating stage indicators and speed ranges into the corresponding thermal balance deduction model parameter set.

[0028] This lookup table pre-defines typical operating stages of refrigerated centrifuges (acceleration stage, constant speed stage, deceleration stage, and standby stage), using the operating stage, speed range, and temperature change gradient as search criteria. It configures corresponding inference model parameter sets and refrigeration compensation coefficients, and can be supplemented with search entries based on actual operating data. For example, when the extracted operating stage marker corresponds to the acceleration stage and the cavity temperature rise rate exceeds a preset rise rate threshold (e.g., 0.5℃ / second), the lookup table outputs an inference model parameter set including the acceleration stage heat load prediction coefficient and pre-cooling compensation amplitude; when the operating stage marker corresponds to the constant speed stage and the cavity temperature fluctuation amplitude is less than a preset stability threshold (e.g., ±0.2℃), the lookup table outputs an inference model parameter set including the constant temperature stage heat balance correction coefficient and energy-saving operating parameters.

[0029] The feature parsing and matching unit 320 encapsulates the converted inference model parameter set, thermal feature vector and operation stage flag into a standard inference input and transmits it to the heat balance inference unit 210. This enables the heat balance inference unit 210 to perform inference based on the thermal characteristics of the centrifuge's actual operation stage rather than generalized parameters, thereby improving the matching accuracy of heat load prediction and cooling capacity demand calculation.

[0030] 3. Parallel scheduling and process coordination: such as Figure 3 As shown, the processing progress of each temperature control cycle is divided into six status nodes: pending data acquisition, feature analysis in progress, deduction calculation in progress, allocation decision in progress, execution confirmation in progress, and archived. After receiving the acquisition ready signal forwarded by the event listening unit 310, the process coordination unit 330, based on the standard deduction input output by the feature analysis and matching unit 320, concurrently calls: the thermodynamic modeling and prediction module 200 (through its internal heat balance deduction unit 210) to perform heat load prediction and cooling demand calculation; and the cooling output execution module 400 to perform dynamic cooling pre-allocation.

[0031] The process coordination unit 330 continuously monitors the completion status of the two parallel branches. Once the heat load prediction result, cooling demand suggestion, and cooling pre-allocation scheme are all generated, the processing progress is updated to "in execution confirmation," and the cooling output execution module 400 is triggered to execute the actual cooling output and feedback collection. After the cooling output is completed, the process coordination unit 330 updates the processing progress to the "archived" state and links the exception information distribution module 600 (through its internal event encapsulation and notification subunit 610) to perform exception event distribution and operation record collection. After the distribution is completed, the processing progress of the temperature control cycle is marked as archived, completing the closed loop of the entire process.

[0032] 4. Operation Status Tracking: The operation status tracking unit 340 uses time points as indexes to record the operation data and dwell time of each sampling moment in the multi-dimensional thermal state perception, thermodynamic modeling, and cooling capacity allocation stages, forming a temperature control decision operation log. This log allows operators or management to query the temperature control status and historical execution time at any time, avoiding the problem of untraceable progress caused by parallel processing of multiple cycle temperature control tasks.

[0033] 5. The fault diagnosis and flexible recovery unit 350 is used to handle situations such as response timeouts, temperature deviations exceeding thresholds, or abnormal sensor data that occur when the process coordination unit 330 calls the thermodynamic modeling and prediction module 200, the refrigeration output execution module 400, the temperature control record generation module 500, or the abnormal information distribution module 600. When a calling link does not return a valid response within a preset timeout period (e.g., 3 seconds), the fault diagnosis and flexible recovery unit 350 reallocates the cooling capacity using a graded derating method according to a preset flexible recovery strategy. For example: When the thermal balance simulation unit 210 fails to return the simulation result within 3 seconds due to excessive computational load, the fault diagnosis and flexible recovery unit 350 automatically switches to the backup simplified simulation model to continue the simulation. If the flexible recovery is successful, the fault diagnosis and flexible recovery unit 350 returns the result to the process coordination unit 330, which continues to advance the processing progress of the temperature control cycle according to the original state node sequence. If all flexible recovery strategies are exhausted and still unsuccessful, the fault diagnosis and flexible recovery unit 350 marks the temperature control cycle as abnormal and requires maintenance, generates an abnormal alarm event, and distributes it to the designated management terminal via the abnormal information distribution module 600 to avoid the failure of the entire machine's temperature control due to a single abnormal temperature control cycle.

[0034] 6. The local edge computing unit 360 deploys all intermediate data storage involved in the thermal balance simulation calculation service, multi-dimensional sensor data parsing service, and process coordination unit 330 within the centrifuge's local embedded controller environment. Sensor data, simulation requests, cooling allocation results, and operational status logs are all stored and processed locally, without being transmitted via public networks, thus ensuring data security and low-latency response requirements in the pharmaceutical and biological testing fields.

[0035] 7. The experience self-learning unit 370 receives the manually reviewed and annotated data returned by the feedback collection subunit 620. It then correlates and matches this data with the original thermal characteristic parameters, projection results, and refrigeration execution results for the corresponding temperature control cycle. It identifies temperature control cases marked as having large prediction deviations or requiring correction, and stores the operating stage markers, thermal characteristic parameter sets, and manual correction results of these cases in a correction record lookup table. For model parameter sets marked as having large prediction deviations, the experience self-learning unit 370 reduces their matching weight in the lookup table. For newly added parameter combinations confirmed as valid after manual correction, the experience self-learning unit 370 adds them to the lookup table and assigns them initial matching weights for priority matching in subsequent similar operating stages and speed ranges. This allows for continuous iterative optimization of the relationship between thermal characteristic types and projection models based on actual operating review data.

[0036] A specific application example of this system: In this embodiment, the sensor data preprocessing and validity verification unit 110 completes the filtering, smoothing, and outlier removal of the raw data from multiple sensors and then hands it over to the integrated sensor data acquisition unit 120. After completing the acquisition of multi-dimensional thermal characteristic data for one temperature control cycle, the integrated sensor data acquisition unit 120 sends an acquisition ready signal, and the event listening unit 310 receives and forwards the signal within 1 second.

[0037] The feature parsing and matching unit 320 analyzes one set of sensor data, identifies the operating stage as the acceleration stage, the current rotor speed as 18,000 rpm, and the cavity temperature rise rate as 0.7℃ / second. Based on the lookup table, it outputs the corresponding acceleration stage heat load prediction coefficient and pre-cooling compensation amplitude and encapsulates them as standard deduction input.

[0038] The process coordination unit 330 synchronously calls the heat balance simulation unit 210 and the cooling output execution module 400. The heat balance simulation unit 210 estimates the heat load change in the next 5 seconds based on the standard simulation input and generates a cooling capacity demand suggestion. The cooling output execution module 400 synchronously generates the corresponding cooling capacity pre-allocation scheme.

[0039] Once both branches are completed, the process coordination unit 330 triggers the cooling output execution module 400 to output the actual cooling capacity command, and links the abnormal information distribution module 600 (through the event encapsulation and notification subunit 610) to distribute the running status and abnormal events to the management terminal.

[0040] If the thermal balance simulation unit 210 fails to return simulation results within 3 seconds due to excessive computational load, the fault diagnosis and flexible recovery unit 350 automatically switches to the backup simplified simulation model. After successful flexible recovery, the results are returned to the process coordination unit 330, and the processing progress continues according to the original state node sequence. The operation status tracking unit 340 records the operation data and dwell time at each state node of the temperature control cycle throughout the entire process for subsequent statistical analysis of temperature control efficiency.

[0041] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A dynamic temperature control system for a refrigerated centrifuge, characterized in that, The system consists of a multi-dimensional thermal state sensing module (100), a thermodynamic modeling and prediction module (200), a temperature control decision module (300), a refrigeration output execution module (400), a temperature control record generation module (500), and an anomaly information distribution module (600); wherein: The multidimensional thermal state sensing module (100) is equipped with a comprehensive sensor data acquisition unit (120). After completing the acquisition of the operating status and thermal parameters, the unit actively broadcasts an acquisition ready signal to the outside. The acquisition ready signal is sent to the event listening unit (310) inside the temperature control decision module (300). After receiving the signal, the event listening unit (310) immediately triggers the process coordination unit (330) configured in the temperature control decision module (300) to enter the working state. The thermodynamic modeling and prediction module (200) is equipped with a thermal balance deduction unit (210). The feature analysis and matching unit (320) deployed in the temperature control decision module (300) obtains output data from the integrated sensor data acquisition unit (120), and obtains multi-dimensional thermal feature indicators by solving the output data. Then, the feature analysis and matching unit (320) converts the multi-dimensional thermal feature indicators obtained by solving into a thermal balance deduction input vector according to the preset thermal feature type and deduction model comparison table, and sends the input vector to the thermal balance deduction unit (210). After the process coordination unit (330) is started, it executes two parallel tasks simultaneously: first, it calls the heat balance simulation unit (210) to perform heat load estimation and cooling demand calculation; second, it calls the cooling output execution module (400) to perform dynamic cooling pre-allocation. After the heat load estimation, cooling demand calculation and dynamic cooling pre-allocation are completed, the process coordination unit (330) triggers the temperature control record generation module (500) to perform the operation record collection operation. After the temperature control record generation module (500) completes the operation record collection, it links the abnormal information distribution module (600) to perform the abnormal event distribution operation.

2. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The multidimensional thermal characteristic indicators collected by the integrated sensor data acquisition unit (120) include at least the current rotor speed, multi-point temperature distribution of the centrifugal cavity, external ambient temperature, real-time operating frequency of the compressor, pressure value of the refrigeration circuit, and the current operating stage indicator; the thermal characteristic type and deduction model comparison table uses the operating stage, speed range, and temperature change gradient as search conditions to configure the corresponding thermal balance deduction model parameter set.

3. The dynamic temperature control system for a refrigerated centrifuge according to claim 2, characterized in that, When the current rotor speed is in the acceleration phase and the cavity temperature rise rate exceeds the preset rise rate threshold, the lookup table output includes characteristic parameters such as the thermal load prediction coefficient and pre-cooling compensation amplitude for the acceleration phase; when the current rotor speed is in the constant speed stable phase and the cavity temperature fluctuation amplitude is less than the preset stability threshold, the lookup table output includes characteristic parameters such as the thermal balance correction coefficient and energy-saving operation parameters for the constant temperature phase.

4. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The process coordination unit (330) continuously monitors the completion status of the two parallel tasks, and updates the working status to "in execution" after the heat load estimation, cooling demand calculation and cooling pre-allocation are completed, triggering the cooling output execution module (400) to output dynamic cooling instructions; after the cooling allocation is completed, the working status is updated to "pending archiving", and the abnormal information distribution module (600) is linked to perform abnormal event distribution.

5. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The temperature control decision module (300) also includes an operation status tracking unit (340). The operation status tracking unit (340) uses time points as indexes to record the operation data and dwell time of each sampling moment in the multi-dimensional thermal state perception, thermodynamic modeling, and cooling capacity allocation stages, forming an operation log for backtracking the temperature control decision process.

6. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The temperature control decision module (300) also includes a fault diagnosis and flexible recovery unit (350), which is used to redistribute the cooling capacity in a graded reduction manner according to a preset flexible recovery strategy when the process coordination unit (330) calls the cooling output execution module (400) and there is a response timeout, temperature deviation exceeds the threshold, or sensor data is abnormal; if the flexible recovery is successful, the result is returned. If elastic recovery fails, it is marked as abnormal and awaits repair, and an alarm event is generated and distributed.

7. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The temperature control decision module (300) also includes an experience self-learning unit (370), and the abnormal information distribution module (600) is provided with an event encapsulation and notification subunit (610) and a feedback collection subunit (620). The experience self-learning unit (370) receives the manually reviewed and labeled data returned by the feedback collection subunit (620) and associates and matches the original thermal feature parameters. It reduces the matching weight of the inference model parameters with large prediction deviations after review and confirmation, and adds the parameter combinations that are confirmed to be effective after manual correction to the lookup table and assigns them initial weights.

8. The dynamic temperature control system for a refrigerated centrifuge according to claim 1, characterized in that, The temperature control decision module (300) also includes a local edge computing unit (360), which is used to deploy the heat balance calculation service, multi-dimensional sensor data parsing service and intermediate data in the scheduling process in the local embedded controller environment of the centrifuge, and isolate the data interaction of the external public network.

9. A dynamic temperature control system for a refrigerated centrifuge according to claim 5, characterized in that, The multi-dimensional thermal state sensing module (100) is also equipped with a sensor data preprocessing and validity verification unit (110), which is used to filter and smooth the original data from multiple sensors and remove outliers before handing it over to the integrated sensor data acquisition unit (120) for feature extraction; the operation status tracking unit (340) divides the status of each temperature control cycle into six status nodes for tracking: pending acquisition, feature analysis in progress, deduction calculation in progress, allocation decision in progress, execution confirmation in progress, and archived.

10. A dynamic temperature control system for a refrigerated centrifuge according to claim 2, characterized in that, The thermal feature type and deduction model comparison table covers four operating stages: acceleration stage, constant speed stage, deceleration stage and standby stage. When the feature analysis and matching unit (320) identifies that the temperature rise rate in the acceleration stage exceeds 0.5℃ / second or the temperature fluctuation amplitude in the constant speed stage exceeds ±0.3℃, it activates and outputs the specific deduction model parameter set for the corresponding stage.

Images