A centrifuge temperature control parameter optimization method and system
By using the XGBoost model to identify centrifuge operating conditions and calculate penalty coefficients, the conflict problem of temperature control under different operating conditions of centrifuges was resolved, adaptive optimization was achieved, and the reliability of experimental results and equipment energy efficiency were improved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- QINGDAO JINHUIYUAN ELECTRONICS CO LTD
- Filing Date
- 2025-10-13
- Publication Date
- 2026-05-19
AI Technical Summary
When existing centrifuges are running under different operating conditions, the PID control algorithm cannot simultaneously achieve fast response, temperature control accuracy, and energy consumption optimization, resulting in temperature overshoot or undercooling, which affects the reliability of experimental results and equipment energy consumption.
The XGBoost model is used for real-time operating condition identification, and the weights of temperature accuracy, energy consumption optimization and temperature change suppression are dynamically adjusted. By constructing a comprehensive objective function, the centrifuge temperature control is optimized, the chamber temperature and cooling power are preset, the penalty coefficient is calculated, and adaptive optimization is achieved.
It achieves intelligent, high-precision, and energy-saving temperature control of centrifuges, significantly improving sample safety and equipment lifespan, and reducing energy waste.
Smart Images
Figure CN121069786B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of centrifuge control technology. In particular, it relates to a method and system for optimizing centrifuge temperature control parameters. Background Technology
[0002] Centrifuges are crucial equipment in fields such as biochemistry, pharmaceuticals, and medical testing for sample separation. Their working principle involves a motor driving a rotor to rotate at high speed, generating centrifugal force to separate samples. During this process, the motor coil dissipates Joule heat, and friction between the rotor and air generates a significant amount of heat, causing rapid temperature changes inside the centrifuge chamber. Many biological samples are highly temperature-sensitive, requiring centrifugation to be conducted in a constant low-temperature environment. Temperature overshoot or overcooling can lead to sample denaturation and inactivation, severely impacting the reliability and reproducibility of experimental results.
[0003] Currently, centrifuge temperature control widely employs traditional PID control algorithms. However, this method has inherent drawbacks in centrifuge operation under varying conditions, including: a fixed set of PID parameters cannot simultaneously adapt to the different requirements of the controller during different operating stages such as acceleration, steady-state, and deceleration. During acceleration, severe thermal disturbances require a rapid controller response to prevent overshoot; during steady-state operation, stable control and low energy consumption are required; during deceleration, heat generation decreases, and overcooling must be prevented, making it difficult for the PID controller to achieve optimal performance in all stages; and the lack of priority switching, focusing solely on minimizing temperature deviation, easily leads to problems such as excessive energy consumption during steady-state operation and excessive cooling during deceleration. Summary of the Invention
[0004] To address the problem that existing centrifuges, when operating under different conditions, cannot accommodate the requirements of various operating conditions, leading to overcooling or overshooting of the centrifuge and affecting the reliability of experimental results, this invention provides solutions in the following aspects.
[0005] In the first aspect, a method for optimizing centrifuge temperature control parameters includes: training an XGBoost stage classification model based on a sample dataset; selecting a target XGBoost model based on recall; and outputting the probability of the centrifuge being in three operating conditions: acceleration, steady speed, and deceleration, based on the target XGBoost model; pre-setting the centrifuge chamber temperature and compressor cooling power for the next moment; calculating a temperature accuracy penalty coefficient based on the deviation between the chamber temperature at the next moment and the preset temperature; calculating a temperature change rate risk coefficient based on the chamber temperature at the next moment and the current moment; calculating an energy consumption penalty coefficient based on the cooling power at the next moment and the maximum value of cooling power in historical data; dynamically adjusting the weights of the temperature accuracy penalty coefficient, temperature change rate risk coefficient, and energy consumption penalty coefficient based on the proportion of steady speed, acceleration, and deceleration in real-time operating conditions and constructing an objective function; calculating the comprehensive control error of the centrifuge at the next moment based on the objective function; selecting multiple sets of historical data most similar to the current operating conditions as reference data; iterating through each reference data set to select the compressor speed corresponding to the reference data that minimizes the comprehensive control error as the optimal speed; and issuing this speed to the compressor for execution.
[0006] By collecting multiple parameters of the centrifuge, the operating status of the centrifuge can be comprehensively reflected. The XGBoost model is used to perform real-time and accurate probabilistic identification of the centrifuge's three operating conditions: acceleration, steady-state, and deceleration. The weights of temperature accuracy, energy consumption optimization, and temperature change suppression are dynamically adjusted according to the proportion of steady-state, acceleration, and deceleration in real-time conditions. For example, more attention is paid to temperature accuracy and energy consumption during the steady-state phase, and more attention is paid to suppressing temperature changes during the acceleration / deceleration phase, thus achieving adaptive optimization. By constructing a comprehensive objective function that includes temperature accuracy penalty, temperature change rate risk, and energy consumption penalty, multiple control objectives that may have conflicted, such as rapid cooling, small temperature fluctuations, and energy saving, are transformed into a single quantifiable optimization objective. By dynamically balancing the sub-objectives through weight coefficients, multi-objective collaborative optimization is achieved, rather than optimization of a single objective. By preset the chamber temperature and cooling power for the next moment and calculating the corresponding penalty coefficient, temperature overshoot, drastic fluctuations, or energy consumption surges can be suppressed in advance, rather than passively responding to errors. This achieves intelligent, high-precision, energy-saving, and adaptive centrifuge temperature control, significantly improving sample safety, equipment lifespan, and operating economy.
[0007] Preferably, the probability of steady speed in the real-time operating condition of the centrifuge is multiplied by 2 and added to the probability of speed increase and speed decrease as the denominator. The probability of steady speed is used as the numerator to calculate the proportion of steady speed in the real-time operating condition, thereby obtaining the temperature accuracy weight and energy consumption optimization weight, which are used as the weights of the temperature accuracy penalty coefficient and the energy consumption penalty coefficient.
[0008] Preferably, the probability of steady speed in the real-time operating condition of the centrifuge is multiplied by 2 and added to the probability of speed increase and speed decrease as the denominator. The sum of the probability of speed increase and speed decrease is used as the numerator to calculate the proportion of the sum of speed increase and speed decrease in the real-time operating condition, thereby obtaining the temperature change suppression weight and using it as the weight of the temperature change rate risk coefficient.
[0009] By multiplying the steady-speed probability by 2, the proportion of the steady-speed probability in the denominator is increased, thereby increasing the weight of temperature accuracy and greatly improving temperature control accuracy. This ensures the safety of sample centrifugation operations. During the acceleration / deceleration phase, by strengthening the weight of temperature change suppression, the rate of temperature change is significantly reduced, avoiding sample damage caused by sudden temperature changes. Under steady-speed or low-load conditions, by increasing the weight of energy consumption optimization, the compressor speed is actively reduced to meet the minimum requirements for temperature accuracy, avoiding energy waste caused by traditional "overcooling" or "over-cooling". Long-term operation can significantly reduce the energy consumption of the centrifuge.
[0010] Preferably, the temperature accuracy penalty coefficient is calculated as follows: the square of the temperature difference between the centrifuge chamber at the next moment and the set temperature is used as the temperature deviation degree, and the ratio of the temperature deviation degree to the maximum value of the temperature deviation degree in historical data is used as the temperature accuracy penalty coefficient.
[0011] Preferably, the temperature change rate risk coefficient is calculated as follows: calculate the rate of change of the temperature of the centrifuge cavity at the next moment compared with the current moment within a sampling period, and use it as the temperature change rate. Calculate the ratio of the temperature change rate to the maximum temperature change rate in the historical data as the temperature change rate risk coefficient.
[0012] Preferably, the energy consumption penalty coefficient is calculated as follows: the ratio of the compressor's cooling power at the next moment to the maximum cooling power in the historical data is used as the energy consumption penalty coefficient.
[0013] Preferably, the method for selecting reference data is as follows: based on the probability of each working condition at the current moment, the working condition with the highest probability is taken as the dominant working condition, all historical data are selected that are the same as the dominant working condition at the current moment, and the Euclidean distance between each selected historical data and the data at the current moment is calculated. Multiple historical data are selected as reference data based on the Euclidean distance.
[0014] First, the historical data is divided into three parts: rising / stable / falling according to the "dominant operating condition" to avoid cross-stage interference. Then, Euclidean distance is used to retain only the multiple samples that are most similar to the current moment, and extreme or abnormal operating conditions are excluded. Since the reference data is highly similar to the current operating condition, the mechanism simulation interpolation error is smaller, which effectively improves the accuracy of the comprehensive control error score and significantly improves the optimal speed prediction accuracy.
[0015] In a second aspect, a centrifuge temperature control parameter optimization system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the centrifuge temperature control parameter optimization method described in any one of the claims is implemented.
[0016] The present invention has the following effects:
[0017] 1. By collecting multiple parameters of the centrifuge, the operating status of the centrifuge can be comprehensively reflected. The XGBoost model is used to perform real-time and accurate probabilistic identification of the centrifuge's three operating conditions: acceleration, steady-state, and deceleration. The weights of temperature accuracy, energy consumption optimization, and temperature change suppression are dynamically adjusted according to the proportion of steady-state, acceleration, and deceleration in real-time conditions. For example, more attention is paid to temperature accuracy and energy consumption during the steady-state phase, and more attention is paid to suppressing temperature changes during the acceleration / deceleration phase, thus achieving adaptive optimization. By constructing a comprehensive objective function that includes temperature accuracy penalty, temperature change rate risk, and energy consumption penalty, multiple control objectives that may have conflicted, such as rapid cooling, small temperature fluctuations, and energy saving, are transformed into a single quantifiable optimization objective. The sub-objectives are dynamically balanced by weight coefficients, achieving multi-objective collaborative optimization rather than optimization of a single objective. By preset the chamber temperature and cooling power for the next moment and calculating the corresponding penalty coefficient, temperature overshoot, drastic fluctuations, or energy consumption surges can be suppressed in advance, rather than passively responding to errors. This achieves intelligent, high-precision, energy-saving, and adaptive centrifuge temperature control, significantly improving sample safety, equipment lifespan, and operating economy.
[0018] 2. During the acceleration / deceleration phase, by strengthening the temperature change suppression weight, the temperature change rate is significantly reduced, avoiding sample damage caused by sudden temperature changes. Under steady speed or low load conditions, by increasing the energy consumption optimization weight, the compressor speed is actively reduced to meet the minimum requirements for temperature accuracy, avoiding energy waste caused by traditional "overcooling" or "over-cooling". Long-term operation can significantly reduce the energy consumption of the centrifuge. Attached Figure Description
[0019] Figure 1 This is a flowchart of steps S1-S4 in a centrifuge temperature control parameter optimization method according to an embodiment of the present invention.
[0020] Figure 2 This is a structural flowchart of a centrifuge temperature control parameter optimization system according to an embodiment of the present invention. Detailed Implementation
[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.
[0022] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0023] Reference Figure 1 A method for optimizing centrifuge temperature control parameters includes steps S1-S4, as detailed below:
[0024] S1: Collect multiple parameters of the centrifuge, construct sample data containing each parameter of the centrifuge based on time-series features, and construct a sample dataset.
[0025] High-precision online sensors and data acquisition cards are used to synchronously collect centrifuge rotor speed, chamber temperature, compressor speed and compressor cooling power. The parameters to be collected also include the preset temperature of the chamber, which is set manually. The acquisition frequency of each parameter can be once per second, and a total of 1,000 complete operating cycles of 50 centrifuges are collected.
[0026] The collected data were preprocessed, and outliers were removed using the 3σ criterion. When there were fewer than 5 consecutive missing sampling points, linear interpolation between adjacent time points was used to fill in the missing points. When there were more than 5 consecutive missing points, the data was marked as invalid cycles. The data were manually labeled into three operating conditions: speed-up, speed-steady, and speed-down, according to the rate of change of the compressor speed.
[0027] Based on time-series features, sample data for centrifuges were constructed, including centrifuge rotor speed, cavity temperature, set temperature, compressor speed, and compressor refrigeration power. A sample dataset was then constructed based on each sample data.
[0028] S2: Train an XGBoost stage classification model based on the sample dataset, select a target XGBoost model based on the recall rate, and output the probability of the centrifuge being in three operating conditions: acceleration, steady speed, and deceleration based on the target XGBoost model.
[0029] Based on the preprocessed sample dataset, a sliding window method is used for feature construction. The step size of the sliding window can be 10 seconds. Taking the current time as the reference, a five-dimensional feature vector is constructed based on the centrifuge rotor speed, cavity temperature, set temperature, compressor speed, and compressor cooling power. The five-dimensional feature vector specifically includes the centrifuge rotor's average speed over 10 seconds, the instantaneous rate of change of the centrifuge rotor speed, the deviation between the actual temperature and the set temperature, the temperature change trend, and the average value of the compressor's cooling power over 10 seconds.
[0030] For each sampling time point, a unified five-dimensional feature vector can be generated according to the above definition. This feature vector is strictly aligned with the operational stage label corresponding to that time point, ensuring complete consistency between features and labels in the time dimension. This lays a reliable data foundation for the accurate training and online recognition of the subsequent model. In other implementations, more statistical features (variance, slope, spectral features, etc.) can be incorporated according to actual needs to further improve the adaptability to complex working conditions.
[0031] The feature vector dataset was divided into training and testing sets in an 8:2 ratio, chronologically, with the first 80% used for training and the last 20% for testing. Grid search and five-fold cross-validation were used to determine the optimal hyperparameter combination for XGBoost (eXtremeGradient Boosting). During model training and parameter optimization, recall for each operating condition (acceleration, steady-state, and deceleration) was used as the core evaluation metric, prioritizing the model with the best recall performance across these conditions as the target XGBoost model.
[0032] The parameters of the centrifuge are collected in real time, and the probability of the centrifuge being in each operating condition at the current moment is output based on the target XGBoost model.
[0033] S3: Preset the centrifuge chamber temperature and compressor cooling power for the next moment. Calculate the temperature accuracy penalty coefficient based on the deviation between the next moment's chamber temperature and the preset temperature. Calculate the temperature change rate risk coefficient based on the chamber temperature at the next moment and the current moment. Calculate the energy consumption penalty coefficient based on the cooling power at the next moment and the maximum value of cooling power in historical data. Dynamically adjust the weights of the temperature accuracy penalty coefficient, temperature change rate risk coefficient, and energy consumption penalty coefficient based on the proportions of steady speed, acceleration, and deceleration in real-time operating conditions, and construct an objective function. Calculate the comprehensive control error of the centrifuge for the next moment based on the objective function.
[0034] The probability of steady speed in real-time centrifuge operation is multiplied by 2 and added to the probabilities of speed increase and decrease as the denominator. The probability of steady speed is used as the numerator to calculate the proportion of steady speed in real-time operation, yielding the temperature accuracy weight and energy consumption optimization weight, which are then used as the weights for the temperature accuracy penalty coefficient and energy consumption penalty coefficient. The probability of steady speed in real-time centrifuge operation is also multiplied by 2 and added to the probabilities of speed increase and decrease as the denominator. The sum of the probabilities of speed increase and decrease is used as the numerator to calculate the proportion of speed increase and decrease in real-time operation, yielding the temperature change suppression weight, which is then used as the weight for the temperature change rate risk coefficient.
[0035] The square of the temperature difference between the centrifuge chamber at the next moment and the set temperature is calculated as the temperature deviation degree. The ratio of the temperature deviation degree to the maximum temperature deviation degree in historical data is calculated and normalized as the temperature accuracy penalty coefficient.
[0036] Calculate the rate of change of the temperature of the centrifuge cavity at the next moment compared to the current moment within a sampling period. Use this as the temperature change rate. Calculate the ratio of the temperature change rate to the maximum temperature change rate in historical data and normalize it to obtain the temperature change rate risk coefficient.
[0037] Calculate the ratio of the compressor's cooling power at the next moment to the maximum cooling power in historical data, and normalize it as an energy consumption penalty coefficient.
[0038] Using temperature accuracy weight as the weight of temperature accuracy penalty coefficient, temperature change suppression weight as the weight of temperature change rate risk coefficient, and energy consumption optimization weight as the weight of energy consumption penalty coefficient, an objective function is constructed to calculate the comprehensive control error of the centrifuge at the next moment.
[0039] The formula for the objective function is: ;
[0040] in, This represents the overall control error of the centrifuge. A smaller value indicates a smaller overall control error, meaning higher temperature accuracy, lower risk of overshoot or overcooling, and better energy efficiency. The purpose of constructing the objective function is to select the compressor speed corresponding to the minimum overall control error at the next moment. Indicates the weight of temperature accuracy. This indicates the temperature accuracy penalty factor. Indicates the temperature change suppression weight. This represents the risk coefficient for temperature change rate. Indicates the weight of energy consumption optimization. This represents the energy consumption penalty coefficient. , , , The value is output in real time based on the target XGBoost model.
[0041] S4: Select multiple sets of historical data that are most similar to the current operating conditions as reference data, iterate through each reference data to select the compressor speed corresponding to the reference data that minimizes the comprehensive control error as the optimal speed, and send it to the compressor for execution.
[0042] When selecting similar historical data, based on the probability of each working condition at the current moment, the working condition with the highest probability is selected as the dominant working condition. Data from all historical data that are the same as the dominant working condition at the current moment are selected, and the Euclidean distance between each selected historical data and the data at the current moment is calculated. The smaller the Euclidean distance, the higher the similarity. Based on the Euclidean distance from smallest to largest, the historical moments are sorted, and the top 20% are selected as reference data. If the number of the top 20% is less than 50, the selection ratio is gradually increased until the number of reference data is greater than 50 to ensure statistical validity. The 50 historical data selected are the most similar to the working condition at the current moment. In other implementations, the number of reference data can be flexibly adjusted as needed.
[0043] The compressor speed data at the next moment corresponding to each reference data point is obtained as candidate speeds. The 95% confidence interval of the candidate speeds is calculated using the normal distribution method, which serves as the effective candidate speed for the current dominant operating condition. This range can cover more than 92% of the effective operating conditions in historical data. Each effective operating condition is successively substituted into the centrifuge and refrigeration system mechanism simulation model, and the cavity temperature and compressor refrigeration power data at the next moment corresponding to each reference data point are output. The temperature accuracy penalty coefficient and temperature change rate risk coefficient are calculated based on the cavity temperature, and the energy consumption penalty coefficient is calculated based on the compressor refrigeration power. Based on real-time data, the operating condition probability is output through the objective XGBoost model, and the temperature accuracy weight, temperature change suppression weight, and energy consumption optimization weight are calculated. The comprehensive control error corresponding to each reference data point is calculated through the objective function after traversing each reference data point.
[0044] The system includes a processor and a memory, the memory storing computer program instructions, which, when executed by the processor, implement the centrifuge temperature control parameter optimization method according to the first aspect of the present invention.
[0045] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.
[0046] The present invention provides a method and system for optimizing centrifuge temperature control parameters, which realizes intelligent, high-precision, energy-saving and adaptive centrifuge temperature control, significantly improving sample safety, equipment life and operating economy.
[0047] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.
Claims
1. A method for optimizing centrifuge temperature control parameters, characterized in that, include: Multiple parameters of the centrifuge were collected, and sample data containing each parameter of the centrifuge was constructed based on time-series features, and a sample dataset was built. An XGBoost stage classification model is trained based on a sample dataset. A target XGBoost model is selected based on recall. The probability of the centrifuge being in three operating conditions—acceleration, steady speed, and deceleration—is output based on the target XGBoost model. Based on the probability of each operating condition at the current moment, the operating condition with the highest probability is selected as the dominant operating condition. Data that is the same as the dominant operating condition at the current moment is selected from all historical data, and the Euclidean distance between each selected historical data and the current data is calculated. Based on the Euclidean distance, multiple historical data are selected as reference data, and the compressor speed data at the next moment corresponding to each reference data is obtained as candidate speed. The normal distribution method is used to obtain the range of effective candidate speeds for the current dominant operating condition. The effective candidate speeds are substituted into the centrifuge and refrigeration system mechanism simulation model, and the cavity temperature and compressor refrigeration power data at the next moment corresponding to each reference data are output. The temperature accuracy penalty coefficient is calculated based on the deviation between the chamber temperature at the next moment and the preset temperature. The temperature change rate risk coefficient is calculated based on the chamber temperature at the next moment and the current moment. The energy consumption penalty coefficient is calculated based on the cooling power at the next moment and the maximum value of cooling power in historical data. The weights of the temperature accuracy penalty coefficient, temperature change rate risk coefficient and energy consumption penalty coefficient are dynamically adjusted based on the proportion of steady speed, speed increase and speed decrease in real-time operating conditions, and an objective function is constructed. The comprehensive control error of the centrifuge at the next moment is calculated based on the objective function. The compressor speed corresponding to the reference data that minimizes the overall control error is selected as the optimal speed and sent to the compressor for execution. Multiply the centrifuge's steady-speed probability by 2 and add it to the acceleration probability and deceleration probability as the denominator. Use the steady-speed probability as the numerator to calculate the proportion of steady-speed in the real-time operating conditions, obtain the temperature accuracy weight and energy consumption optimization weight, and use them as the weights of the temperature accuracy penalty coefficient and energy consumption penalty coefficient. Multiply the centrifuge's steady-speed probability by 2 and add it to the acceleration and deceleration probabilities as the denominator. Use the sum of acceleration and deceleration probabilities as the numerator to calculate the proportion of acceleration and deceleration in the real-time operating conditions. This yields the temperature change suppression weight, which is then used as the weight of the temperature change rate risk coefficient.
2. The method for optimizing centrifuge temperature control parameters according to claim 1, characterized in that, The temperature accuracy penalty coefficient is calculated as follows: The square of the difference between the chamber temperature at the next moment and the set temperature is calculated as the temperature deviation degree, and the ratio of the temperature deviation degree to the maximum temperature deviation degree in historical data is calculated as the temperature accuracy penalty coefficient.
3. The method for optimizing centrifuge temperature control parameters according to claim 1, characterized in that, The calculation method for the temperature change rate risk coefficient is as follows: Calculate the rate of change of the chamber temperature at the next moment within a sampling period compared to the current moment, and use this as the temperature change rate. Calculate the ratio of the temperature change rate to the maximum temperature change rate in historical data as the temperature change rate risk coefficient.
4. The method for optimizing centrifuge temperature control parameters according to claim 1, characterized in that, The energy consumption penalty coefficient is calculated as follows: The ratio of the compressor's cooling power at the next moment to the maximum cooling power in historical data is used as the energy consumption penalty coefficient.
5. A centrifuge temperature control parameter optimization system, characterized in that, include: A processor and a memory, the memory storing computer program instructions that, when executed by the processor, implement the centrifuge temperature control parameter optimization method according to any one of claims 1-4.