An intelligent monitoring method for ultracentrifuge tubes based on sensor network

Abstract

This invention provides a sensor network-based intelligent monitoring method for ultracentrifuge tubes, belonging to the field of ultracentrifuge tube technology. The method comprises: S1. deploying and initializing the sensor network; S2. performing real-time data acquisition and preprocessing; S3. constructing a centrifuge tube physical model and calibrating its parameters; S4. multi-source data fusion and feature extraction; S5. abnormal state identification and early warning decision-making; and S6. closed-loop feedback control and parameter adaptation. By acquiring multi-dimensional data such as centrifuge tube vibration, temperature, internal pressure, and liquid level changes, the method comprehensively covers key status parameters during centrifuge tube operation. Compared with traditional single-sensor monitoring, the method improves the integrity of monitoring information.

Description

Technical Field

[0001] The invention relates to the technical field of ultracentrifuge tubes, and in particular to an intelligent monitoring method for ultracentrifuge tubes based on a sensor network. Background Art

[0002] In fields like biomedicine and chemical analysis, ultracentrifuges are critical equipment for separating and purifying samples such as biomolecules and cellular components. As core consumables, ultracentrifuge tubes must maintain structural integrity and sample stability even at high rotation speeds (up to tens of thousands of revolutions per minute). However, existing monitoring methods have significant shortcomings. Traditional monitoring methods often rely on a single type of sensor (e.g., monitoring only rotational speed or temperature), failing to fully capture the tube's changing state under complex operating conditions. For example, abnormal vibrations caused by liquid leaks and imbalances can be difficult to detect promptly. Existing monitoring methods often employ simple logic such as threshold judgments, lacking the deep integration and analysis of multi-source data. This makes it difficult to accurately identify potential risks and can easily generate false positives or omissions, leading to equipment failure and even sample loss. Furthermore, existing monitoring systems often lack adaptive adjustment capabilities, making it impossible to dynamically optimize operating parameters based on the tube's real-time state, making it difficult to meet the demands of high-precision experiments and production. Summary of the Invention

[0003] The present invention provides an ultracentrifuge tube intelligent monitoring method based on a sensor network to solve one or more of the above problems.

[0004] To achieve the above object, the present invention adopts the following technical solutions:

[0005] A sensor network-based intelligent monitoring method for ultracentrifuge tubes, comprising:

[0006] S1. Deploy and initialize the sensor network. Deploy multiple types of sensors on the fixed base, rotating shaft, inner wall of the centrifuge chamber, and each centrifuge tube to form a three-dimensional distributed monitoring network. These sensors include triaxial accelerometers, temperature sensors, pressure sensors, and liquid level sensors. Connect all sensors to the edge computing gateway via wired or wireless connections.

[0007] S2. Perform real-time data acquisition and preprocessing. The edge computing gateway synchronously collects all sensor data, applies a low-pass filter to the raw acceleration signal to remove high-frequency noise, applies a sliding average filter to the raw temperature signal to eliminate random interference, and performs threshold verification on the raw pressure and liquid level signals to remove outliers.

[0008] S3. Construct a centrifuge tube physical model and calibrate its parameters. Establish a centrifugal force-stress model and a heat conduction model based on the centrifuge tube's specifications. Calibrate the model parameters using historical data. Fit the bracket's vibration transfer function using no-load operation data. Determine the relationship between liquid sloshing and liquid level through load testing.

[0009] S4. Multi-source data fusion and feature extraction: The preprocessed sensor data is spatiotemporally aligned with the centrifugal force-stress model output. Extended Kalman filtering is used to fuse multi-source information and estimate the centrifuge tube's real-time state vector. Feature parameters are extracted from the centrifuge tube's real-time state vector, including speed change rate, temperature rise rate, pressure fluctuation coefficient, vibration energy ratio, liquid level gradient, and centrifugal force deviation rate.

[0010] S5. Abnormal state identification and warning decision-making: Input feature parameters into the random forest classifier to identify the state categories as normal, slight vibration, overheating, overload, and liquid leakage. If it is determined to be abnormal, a level 1 warning, level 2 warning, or level 3 warning is triggered according to preset rules.

[0011] S6. Closed-loop feedback control and parameter adaptation; if a level 1 or level 2 warning is detected, the edge computing gateway sends a control instruction to the centrifugal device to adjust the motor speed through the PID controller; at the same time, the parameters of the bracket vibration transfer function, centrifugal force-stress model, or heat conduction model are dynamically updated based on real-time monitoring data.

[0012] In this manual, the three-axis acceleration sensor is attached to the central axis of the centrifuge tube side wall, the temperature sensor is embedded in the bottom of the centrifuge tube, the pressure sensor is integrated into the inside of the tube cover, and the liquid level sensor is fixed at 1 / 2 the height of the tube body using an infrared tube structure.

[0013] In this specification, the low-pass filter used for the original acceleration signal is a Butterworth low-pass filter.

[0014] In this specification, the centrifugal force-stress model is constructed based on the centrifugal force formula, and the heat conduction model is established according to the specific heat capacity and heat dissipation coefficient of the tube material.

[0015] In this specification, the state transfer function of the extended Kalman filter is determined according to the change law of the centrifuge tube speed, and the process noise covariance matrix is obtained through sensor error statistics.

[0016] In this specification, the training samples of the random forest classifier include normal operating condition data and various abnormal operating condition historical data.

[0017] In this manual, the first-level warning displays a warning message on the device screen, the second-level warning sends a text message notification to the monitoring terminal and records the abnormality log, and the third-level warning automatically triggers the device shutdown program and cuts off the motor power.

[0018] In this manual, the proportional, integral, and differential coefficients of the PID controller are adjusted using the critical proportionality method.

[0019] In this specification, in S6, when the liquid level changes by more than 5%, the total mass of the centrifuge tube is recalculated, including the mass of the sample liquid and the mass of the tube body.

[0020] In this specification, in S6, by comparing the residuals of the actual monitoring data with the model prediction data, when the residual exceeds a certain threshold, the model parameters are recalibrated.

[0021] In summary, the present invention has at least the following beneficial effects:

[0022] Multi-dimensional precise monitoring: By deploying multiple sensors such as triaxial acceleration, temperature, pressure, and liquid level at key locations on centrifuge equipment and centrifuge tubes, a three-dimensional distributed monitoring network is constructed. This allows real-time acquisition of multi-dimensional data such as centrifuge tube vibration, temperature, internal pressure, and liquid level changes, comprehensively covering the key status parameters of the centrifuge tube during operation. Compared with traditional single-sensor monitoring, this improves the integrity of monitoring information.

[0023] Intelligent anomaly recognition: Integrating algorithms such as non-negative matrix factorization (NMF), radial basis function (RBF) surface fitting, and random forest classification, it performs dimensionality reduction, nonlinear modeling, and feature fusion on multi-source data. It can accurately identify various abnormal conditions, including normal and slight vibration, overheating, overload, and liquid leakage. Compared with traditional threshold methods, it improves the accuracy of anomaly detection and effectively reduces the false alarm and missed alarm rates.

[0024] Dynamic Adaptive Control: Based on anomaly detection results, the system automatically implements multi-level early warning and closed-loop control strategies. During a Level 1 alert, the sampling frequency is increased to closely track anomaly trends; during a Level 2 alert, the PID controller precisely adjusts the speed; and during a Level 3 alert, an emergency shutdown is initiated to prevent equipment damage and sample loss. Simultaneously, physical model parameters and algorithm models are dynamically updated based on real-time data, enabling the system to adapt to different operating conditions and sample loads, shortening control response time.

[0025] A comprehensive data-driven monitoring system encompasses the entire process, from sensor data acquisition, preprocessing, and feature extraction to model building and decision-making control. This system integrates closely linked processes and provides traceable data, creating a comprehensive data-driven monitoring system. This not only improves the safety and stability of the centrifugation process but also provides rich data support for subsequent process optimization and equipment maintenance, extending equipment life and reducing sample loss costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0027] Figure 1 Schematic diagram of the sensor network-based intelligent monitoring method for ultracentrifuge tubes involved in the present invention.

[0028] Figure 2 Schematic diagram of the sensor deployment and initialization process involved in the present invention.

[0029] Figure 3 The figure is a schematic diagram of the abnormal state identification and early warning process involved in the present invention. DETAILED DESCRIPTION

[0030] Hereinafter, only certain exemplary embodiments are briefly described. As will be appreciated by those skilled in the art, the described embodiments may be modified in various ways without departing from the spirit or scope of the embodiments of the present invention. Therefore, the drawings and description are to be regarded as illustrative in nature and not restrictive.

[0031] The disclosure below provides many different embodiments or examples for implementing different structures of the embodiments of the present invention. In order to simplify the disclosure of the embodiments of the present invention, the components and configurations of specific examples are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. In addition, the embodiments of the present invention may repeat reference numerals and / or reference letters in different examples. Such repetition is for the purpose of simplicity and clarity and does not in itself indicate the relationship between the various embodiments and / or configurations discussed.

[0032] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0033] like Figure 1 As shown, this embodiment provides an ultracentrifuge tube intelligent monitoring method based on a sensor network, comprising the following steps:

[0034] S1. Deploy and initialize the sensor network. Deploy a triaxial accelerometer, temperature sensor, pressure sensor, and liquid level sensor inside the ultracentrifuge chamber, on the centrifuge tube holder, and on the outer wall of each centrifuge tube. Connect all sensors to the edge computing gateway via wired or wireless connections, registering hardware addresses and initializing communication protocols.

[0035] S2. Perform real-time data acquisition and preprocessing. The edge computing gateway synchronously collects all sensor data at a set frequency. It applies a low-pass filter to the raw acceleration signal to remove high-frequency noise, a sliding average filter to the raw temperature signal to eliminate random interference, and threshold checks to remove outliers from the raw pressure and liquid level signals.

[0036] S3. Construct a centrifuge tube physical model and calibrate its parameters. A centrifugal force model and a heat conduction model were established based on the centrifuge tube's specifications. Model parameters were calibrated using historical data. The bracket's vibration transfer function was fitted using no-load operating data. The relationship between liquid sloshing and liquid level was determined through load testing. The centrifugal force model was constructed based on the centrifugal force formula, and the heat conduction model was established based on the specific heat capacity and heat dissipation coefficient of the tube material.

[0037] S4. Multi-source data fusion and feature extraction: align the preprocessed sensor data with the centrifugal force model output constructed in S3 in time and space, fuse the multi-source information using an extended Kalman filter, and estimate the real-time state vector of the centrifuge tube, which includes speed, temperature, pressure, liquid level, and peak vibration acceleration.

[0038] The vibration transfer function of the bracket is used to separate the vibration characteristics of the equipment and the centrifuge tube itself;

[0039] Correct the liquid level measurement value according to the relationship between liquid sloshing and liquid level;

[0040] Calculate the theoretical temperature rise rate by combining the heat conduction model;

[0041] The following characteristic parameters are extracted:

[0042] Speed change rate: the ratio of the speed difference between two adjacent sampling moments to the sampling time interval;

[0043] Temperature rise rate: the ratio of the temperature difference between two adjacent sampling moments to the sampling time interval;

[0044] Pressure fluctuation coefficient: the ratio of the standard deviation of pressure data to the average pressure;

[0045] Vibration energy ratio: By performing Fourier transform on the acceleration signal, the ratio of the vibration energy within a specific frequency range to the total vibration energy is calculated;

[0046] Liquid level change gradient: the ratio of the difference between the liquid levels at two adjacent sampling moments to the sampling time interval;

[0047] Centrifugal force deviation rate: the difference between the actual calculated centrifugal force and the average centrifugal force during normal operation, and the ratio of the difference to the average centrifugal force during normal operation;

[0048] S5. Abnormal state identification and early warning decision-making; input the feature parameters extracted in S4 into the random forest classifier to identify the state categories as normal, slight vibration, overheating, overload, and liquid leakage; the process is as follows Figure 3 As shown;

[0049] When the measured temperature rise rate exceeds 120% of the value predicted by the heat conduction model, it is determined to be a heat conduction anomaly. If the heat conduction anomaly causes the temperature rise rate to exceed 50% of the critical temperature rise rate, a second-level warning is triggered, a text message notification is sent to the monitoring terminal, and an abnormality log is recorded. If the second-level warning conditions are not met but the temperature exceeds the normal range, a first-level warning is triggered and a warning message is displayed on the device screen.

[0050] When the equipment vibration component accounts for more than 70% of the vibration characteristics, it is determined that the bracket vibration is abnormal. If the abnormal bracket vibration causes the vibration peak acceleration to exceed 80% of the maximum tolerance value, a second-level warning is triggered, and a text message notification is sent to the monitoring terminal and an abnormality log is recorded. If the second-level warning conditions are not met, but the pressure fluctuation coefficient exceeds the normal range by 10%-20%, a first-level warning is triggered and a warning message is displayed on the device screen.

[0051] When the speed change rate exceeds the set threshold 1 and does not exceed threshold 2 (which can be set according to actual conditions, for example, threshold 1 is set to 5% / s (speed change exceeds 5% per second) and threshold 2 is set to 10% / s), or the pressure fluctuation coefficient exceeds the normal range of 10%-20%, it is determined to be a slight vibration abnormality, triggering a level 1 warning and displaying a warning message on the device screen;

[0052] When the temperature rise rate exceeds 50% of the critical temperature rise rate, or the centrifugal force reaches 80%-90% of the maximum tolerable centrifugal force, it is judged as overheating or overload abnormality, triggering a secondary warning, sending a text message notification to the monitoring terminal and recording the abnormality log;

[0053] When the liquid level drops by more than 30%, or the vibration peak acceleration exceeds the maximum tolerance value, it is determined to be a liquid leak or severe vibration abnormality, triggering a three-level warning, automatically triggering the equipment shutdown program, and cutting off the motor power supply;

[0054] S6. Closed-loop feedback control and parameter adaptation. If an adjustable anomaly (other than a Level 3 warning) is detected, the edge computing gateway sends a control command to the centrifuge device, adjusting the motor speed via a PID controller. Simultaneously, the centrifugal force model and heat conduction model parameters in S3 are dynamically updated based on real-time monitoring data. When the liquid level changes by more than 5%, the total mass of the centrifuge tube, including the sample liquid mass and the tube mass, is recalculated.

[0055] When the abnormal vibration of the bracket lasts for more than 5 minutes, recalibrate the bracket vibration transfer function;

[0056] When the heat conduction anomaly occurs three times in a row, the heat dissipation coefficient parameters in the heat conduction model are updated.

[0057] In some embodiments, as Figure 2 As shown in the figure, to comprehensively monitor the status of the centrifuge tubes during operation, multiple types of sensors must be deployed at key locations. First, within the ultracentrifuge chamber, this area serves as the core operating space for the centrifuge tubes and directly reflects the overall operating environment of the equipment. Sensors are deployed here to monitor environmental parameters. Sensors are also deployed on the centrifuge tube holders, as the holder's condition directly affects the stability of the tubes. Sensors are also deployed on the outer walls of each tube to obtain the tube's own operating parameters. Specifically, these sensors include a triaxial accelerometer to monitor the tube's vibration during operation; a temperature sensor to monitor temperature changes in real time; a pressure sensor to measure the internal pressure of the tube; and a liquid level sensor to detect the liquid level within the tube. After deployment, all sensors are connected to the edge computing gateway via wired (e.g., Ethernet) or wireless (e.g., ZigBee) connections. Hardware address registration is performed, assigning each sensor a unique identification address for subsequent data transmission and management. Communication protocol initialization is also completed, determining parameters such as the data transmission format and frequency between the sensor and the gateway to ensure accurate and stable data transmission.

[0058] In some embodiments, when installing the three-axis acceleration sensor, it is precisely pasted on the center axis position of the side wall of the centrifuge tube, which can most accurately reflect the vibration state of the centrifuge tube; the temperature sensor is embedded in the bottom of the centrifuge tube to ensure that the temperature changes inside the centrifuge tube can be sensed in a timely and accurate manner; the pressure sensor is integrated on the inside of the tube cover to directly measure the internal pressure of the centrifuge tube; the liquid level sensor uses an infrared tube structure to be fixed at 1 / 2 the height of the tube body for real-time monitoring of the liquid level. During the hardware address registration process, the edge computing gateway sends a registration instruction to each sensor. After receiving the instruction, the sensor feeds back its own hardware address information to the gateway, and the gateway records and manages the address. When the communication protocol is initialized, the gateway negotiates with the sensor to determine the sampling frequency, for example, set it to 100Hz, that is, collect data 100 times per second; determine the data transmission format as JSON format, including fields such as sensor ID, timestamp, physical quantity type, and numerical value to ensure the standardization and readability of the data.

[0059] In some embodiments, the edge computing gateway synchronously collects data from all sensors at a set frequency, such as 100Hz. Since the raw data collected by the sensor may contain noise, interference or outliers, it needs to be preprocessed. For the original acceleration signal, it may contain high-frequency noise, which will affect the judgment of the true vibration state of the centrifuge tube. Therefore, a Butterworth low-pass filter is used to remove high-frequency noise to make the acceleration signal smoother and accurately reflect the vibration of the centrifuge tube. The original temperature signal may have random interference. It is processed by sliding average filtering to eliminate these random fluctuations and obtain more stable temperature data. The original pressure and liquid level signals may have outliers that exceed the normal range. These outliers are eliminated by threshold verification to ensure the validity of the pressure and liquid level data. After the above processing, the processed data is organized into a multi-sensor data matrix , providing a basis for subsequent data processing and analysis.

[0060] In some embodiments, when performing Butterworth low-pass filtering, the original acceleration signal is first Substitute the digital filter formula obtained by converting the analog filter through bilinear transformation:

[0061] ;

[0062] in is the transfer function of the digital filter, which describes the processing ability of the digital filter for signals of different frequencies. Among them, z is the variable in the discrete signal Z transform, which is used to represent the frequency domain characteristics of the discrete time series. The coefficient =0.043, =0.086, =0.043, =-1.226, =0.498 (cut-off frequency 50Hz, second order). According to this formula, the acceleration data of each sampling point is calculated to obtain the filtered acceleration signal , effectively remove high-frequency noise. :It is the numerator of the filter transfer function The coefficient of the term has a direct impact on the gain of the filter at zero frequency (DC component). In a low-pass filter, it determines the degree of preservation of the DC component (low-frequency stable part) in the signal. For example, when When it is larger, it means that the filter has a stronger amplification effect on the DC component; on the contrary, when When it is small, the DC component will be relatively weakened. and :These two coefficients are in the numerator and The coefficients of the terms, together with the denominator coefficients, determine the shape of the filter’s phase-frequency and amplitude-frequency characteristics. and The value of can change the filter's attenuation speed and phase delay for signals of different frequencies. In the Butterworth low-pass filter design, they work together with the denominator coefficient to make the filter have a flat amplitude-frequency response in the passband and achieve fast attenuation in the stopband. Taking the second-order Butterworth low-pass filter as an example, the appropriate and The value can ensure that the signal transitions from the passband to the stopband is smoother near the cutoff frequency. and :They are the denominators and The coefficient of the term plays a key role in determining the filter characteristics. and The denominator coefficient and the numerator coefficient directly affect the stability and frequency response of the filter. For the Butterworth low-pass filter, the value of the denominator coefficient determines the order and cutoff frequency of the filter. In the second-order Butterworth low-pass filter, the specific and Value (such as =-1.226, =0.498, corresponding to a cutoff frequency of 50Hz) ensures maximum filter flatness within the passband and attenuates high-frequency signals at a predetermined rate starting at the cutoff frequency. This ensures stable filter operation at all frequencies, avoiding instabilities such as oscillation. The values of these coefficients are determined based on the filter's design requirements (such as cutoff frequency, filter order, passband ripple, and stopband attenuation) through mathematical calculations (such as bilinear transformation and impulse response invariance). Taking the bilinear transformation approach to designing a second-order Butterworth low-pass filter as an example, the analog filter parameters are first calculated based on the desired cutoff frequency. Then, the analog filter's transfer function is converted to a digital filter's transfer function through a bilinear transformation, resulting in the coefficient values for the digital filter's transfer function.

[0063] In some embodiments, for the temperature signal, the temperature sequence is read in chronological order. , for every M1=5 consecutive data points, use the formula Calculate the sliding average to obtain the smoothed temperature series , eliminating random interference.

[0064] In some embodiments, when processing pressure and liquid level signals, the real-time pressure And level data and preset range Compare. If the pressure Out of range, that is, not satisfied , then take the adjacent effective values for linear interpolation, according to the formula Calculate the reasonable pressure value and process the liquid level data in the same way. Finally, align the processed acceleration, temperature, pressure and liquid level data in chronological order to form a multi-sensor data matrix , where M is the number of samples and n=4 represents the number of sensor types.

[0065] In some embodiments, in order to accurately analyze the physical state of the centrifuge tube during operation, it is necessary to determine the material density of the centrifuge tube. , volume , Maximum tolerable speed The centrifugal force-stress model and heat conduction model are established based on the specification parameters. These models can theoretically describe the physical changes of the centrifuge tube under different working conditions. In order to make the model more consistent with the actual situation, the vibration data of the equipment base and the centrifuge tube are collected through no-load operation experiments, and the vibration transfer function of the bracket is fitted using the least squares method to find the relationship between the vibration of the equipment base and the vibration of the centrifuge tube. Then, the liquid level in the centrifuge tube is changed through load experiments. , monitoring liquid sloshing and determining the relationship between liquid sloshing and liquid level. Finally, using data collected during the actual operation of the equipment, the established model parameters are calibrated, enabling the model to more accurately predict the actual operating status of the centrifuge tube.

[0066] In some embodiments, when establishing the centrifugal force model, according to the formula ,in is the centrifugal force, is the angular velocity, is the rotation radius, and the calculation of the total mass m (of the centrifuge tube and the sample inside) involves the mass of the liquid and the mass of the tube. The mass of the liquid is determined by the density of the liquid. and liquid volume Determine the volume of liquid The mass of the tube is ,so .

[0067] In some embodiments, the heat conduction model is based on the formula Build, where is the rate of change of temperature T with time t, that is, the temperature rise rate of the centrifuge tube, is the frictional heat power, is the specific heat capacity, is the heat dissipation coefficient, is the current temperature, is the ambient temperature.

[0068] In some embodiments, when performing a no-load operation experiment, the ultracentrifuge device is started and operated in a no-load state, and the vibration signal of the device base is collected by a sensor. and centrifuge tube vibration signal , and the Laplace transformed and Substitute the transfer function , use the least squares method to calculate the parameters 、 、 、 For example, =0.5, =0.2, =0.8, =0.1, determine the vibration transfer function of the bracket. In the load experiment, different volumes of liquid were injected into the centrifuge tube to change the liquid level. , operate the equipment and monitor the vibration of the centrifuge tube, and establish the sloshing acceleration through data analysis With liquid level relationship, such as In the actual operation process, the real-time collected temperature, vibration and other data are substituted into the established model, the initial model parameter values are set, and the heat dissipation coefficient is continuously adjusted through the gradient descent method. and frictional heat power , so that the error between the temperature predicted by the model and the actual measured temperature is minimized, completing the calibration of the model parameters.

[0069] The goal of the least squares method is to make the observed data and the model prediction value The sum of squared errors between the two is minimized, and the error function is assumed to be , where n is the number of data samples, By transfer function and The predicted value is calculated. The minimum value of Find the partial derivative and set it to 0:

[0070] ;

[0071] Solving the simultaneous equations yields The value of is used to determine the bracket vibration transfer function.

[0072] In some embodiments, after obtaining the pre-processed multi-sensor data matrix Finally, due to the high dimension of the data and the existence of information redundancy, in order to extract more valuable features, non-negative matrix factorization (NMF) is performed on it. Through the NMF algorithm, the high-dimensional data is mapped to a low-dimensional space to obtain the basis matrix and coefficient matrix (i.e., feature matrix) , thereby extracting the implicit characteristic patterns of the data. Then, the feature matrix obtained by NMF (including speed change rate, temperature rise rate, pressure fluctuation coefficient, vibration energy ratio, liquid level change gradient, centrifugal force deviation rate and other elements) and the physical parameters of the centrifuge tube, such as speed , liquid level As input, radial basis function (RBF) surface fitting technology is used to establish a nonlinear mapping model between sensor data. The model can predict physical quantities such as pressure and vibration of the centrifuge tube based on the input features and physical parameters, and realize the deep fusion and analysis of multi-source data.

[0073] In some embodiments, when performing non-negative matrix decomposition, the multi-sensor data matrix As input, the feature dimension k1=3 is preset (this value is determined by cross-validation to ensure that the data features are retained to the greatest extent while reducing the dimension). First, the basis matrix is randomly initialized and coefficient matrix , and then calculate according to the iterative update formula, the iterative update formula:

[0074] ;

[0075] They are sample index, feature index, and sensor dimension index respectively. In each iteration, they are updated according to the formula and The element value of the matrix is adjusted continuously until the reconstruction error Convergence (the number of iterations is generally set to 500 times), and finally the feature matrix after dimensionality reduction is obtained When constructing the RBF surface model, , speed , liquid level Composing the input vector , the actual measured pressure As output target . Select sample points according to the RBF model formula: ,in For the The weight of each basis function determines the contribution of each basis function to the final prediction result; the number of basis functions =2g, g is the number of samples, basis function center , is the kernel width parameter, is the linear term coefficient. Use the least squares method to solve the model parameters , that is, by adjusting the parameters so that the loss function Minimize and complete the training of RBF model. In the model application stage, the real-time collected 、 、 Composing the input vector , input into the trained RBF model, and calculate the predicted pressure .

[0076] In some embodiments, in order to ensure the accuracy and reliability of the constructed RBF surface model, it is necessary to verify it. The actual measured data is input into the trained RBF surface model, and the residual between the model's predicted value and the true value is calculated. By calculating the mean of the statistical residuals and standard deviation , evaluate the fitting accuracy of the model. If the residual exceeds a certain threshold, that is (Based on the three-times standard deviation criterion, a criterion often used in statistics to determine data abnormality), this is considered an abnormal data point. For these abnormal data points, the system triggers a data resampling mechanism to recollect the data at that moment. If multiple abnormal data points appear consecutively, the model is retrained to ensure that it accurately reflects the actual operating status of the centrifuge tube.

[0077] In some embodiments, when calculating the residual, for each sampling point, the actual measured pressure value is Predicted pressure value with RBF model Substitute into the formula , and get the residual value As data is continuously collected, the mean of the residuals is calculated using a sliding window (the window size can be set according to the actual situation, such as including 10 sample points) and standard deviation , the calculation formula is:

[0078] ;

[0079] After each calculation of the residual statistics, each residual value With threshold For comparison, if , the data point is determined to be an anomaly, and the edge computing gateway sends a command to the sensor to recollect the relevant data at that moment. If an anomaly is detected in multiple consecutive sliding windows (for example, three windows), the model is considered to have deviated, triggering the RBF model retraining process. The model is retrained using the latest collected data and the model parameters are adjusted to improve its accuracy.

[0080] In some embodiments, after extracting the NMF feature matrix And calculate the surface fitting residual , residual mean , residual standard deviation and the maximum residual value Then, these data are fused into a new feature vector The feature vector is input into the improved random forest classifier, and the powerful classification ability of the random forest classifier is used to identify the status of the centrifuge tube. The output status categories include normal, slight vibration, overheating, overload, liquid leakage and model mismatch. In the recognition process, according to the residual standard deviation threshold (such as ,in is the standard deviation of the training set residuals) to determine whether to trigger the model adaptive update, so as to adjust the model in time to ensure the accuracy and reliability of the monitoring system.

[0081] In some embodiments, first, the NMF feature matrix , residual mean , residual standard deviation and the maximum residual value Combine them in order to form a new feature vector When training a random forest classifier, prepare a large amount of historical feature vector data The corresponding accurate state labels (normal, slight vibration, etc.) are used as training samples. The parameters of the random forest are set to build a classification model containing 100 decision trees. When training each decision tree, 70% of the data and 50% of the features are randomly sampled from the training samples for training. This method increases the diversity and generalization ability of the model. After the training is completed, the feature vector generated in real time is Input into the trained random forest classifier. Each decision tree in the classifier classifies the input feature vector and obtains a classification result. The final classification result is determined by the voting mechanism, namely:

[0082] ;

[0083] in For the A decision tree, is the indicator function, represents the first In the application scenario of ultracentrifuge tube intelligent monitoring, these classification categories correspond to different possible operating states of the centrifuge tube, specifically: : Indicates "normal" status, that is, the centrifuge tube operating parameters are within the normal range and no abnormalities occur; : Indicates a "mild vibration" abnormal state, which means that the vibration amplitude of the centrifuge tube exceeds the normal range but has not yet reached a serious level; : Indicates an "overheat" abnormal state, that is, the temperature of the centrifuge tube is too high, which may affect the normal operation of the equipment and the quality of the sample; : Indicates the "overload" abnormal state, indicating that the load borne by the centrifuge tube exceeds its designed bearing range; : Indicates the abnormal state of "liquid leakage", indicating that the liquid in the centrifuge tube is leaking; : Indicates the abnormal state of "model mismatch", that is, the residual of the surface fitting model exceeds the expected range and the model cannot accurately reflect the actual situation.

[0084] when hour is 1, otherwise it is 0. After obtaining the classification results, the current residual standard deviation The standard deviation of the residuals from the training set For comparison, if , it is determined to be in a "model mismatch" state, triggering the subsequent model adaptive update process. If it is determined to be in another abnormal state (such as slight vibration or overheating), the corresponding level of warning will be triggered according to the preset warning rules. For example, if it is identified as a slight vibration, the device's local indicator light may flash as a prompt. If it is a more serious abnormality such as overheating or overload, not only will an audible and visual alarm be issued locally, but detailed abnormality information including the abnormality type, occurrence time, and relevant characteristic parameters will be sent to the remote monitoring terminal to enable operators to take timely measures. At the same time, the system records the results and related data of each abnormality identification for subsequent data analysis and model optimization.

[0085] In some embodiments, after the random forest classifier outputs an anomaly category, the system triggers an alert based on a pre-set anomaly-alert mapping table. This mapping table clearly defines the alert levels and handling methods corresponding to different anomaly categories, for example:

[0086] Exception Category Warning level Treatment Slight vibration Level 1 warning <![CDATA[The local yellow indicator light of the device flashes at a frequency of 1 time per second, and the prompt message "轻微振动异常 (Minor vibration anomaly)" is displayed on the device operation interface for 10 minutes; the abnormal information is recorded in the local log file, including the timestamp, X new feature vector data]]> overheat Level 2 warning The device emits a continuous beeping sound and the red indicator light stays on. A text message and push notification containing the message "Overheating abnormality, current temperature: [measured temperature value], it is recommended to check the cooling system immediately" is sent to the remote monitoring terminal. An abnormal work order is generated on the device management platform and assigned to maintenance personnel. Overload Level 2 warning Same as overheating treatment method, detailed overload data (current centrifugal force, speed, etc.) is displayed on the equipment operation interface at the same time Liquid leakage Level 3 warning The device will emit a shrill alarm and the red indicator light will flash rapidly (5 times / second). The device will automatically cut off the power supply and stop the centrifugal operation. An emergency notification will be sent to all relevant personnel (operators, managers, and maintenance personnel) containing the message "Liquid leak, the device has stopped, please go to the scene immediately." The location of the leaking device and related data will be highlighted on the monitoring screen.

[0087] When recording abnormal data, the system will record the current time, abnormal category, feature vector The information such as the predicted value and measured value of the RBF model is stored in the database in a specific format (such as JSON format), which is convenient for the subsequent analysis of the abnormal occurrence pattern through data mining technology, and provides a basis for further optimizing the random forest model and adjusting the early warning strategy.

[0088] In some embodiments, when S5 detects an adjustable abnormality (such as minor vibration, overheating and other non-serious emergencies other than the third-level warning), the edge computing gateway adjusts the motor speed through the PID controller according to the preset control strategy to try to restore the centrifuge tube to normal operation. When key parameters such as tube mass change, the system will use the RBF surface model to infer the unknown parameters of the centrifuge tube, such as , heat dissipation coefficient The objective function is continuously optimized through the gradient descent method to minimize the error between the model prediction value and the measured value, realize the online calibration of the physical model parameters, and then update the entire monitoring model to adapt to the changes in the operating state of the centrifuge tube, forming a complete closed-loop feedback control system.

[0089] In some embodiments, the PID control process: If an adjustable abnormality such as slight vibration or overheating is detected, the system obtains the current speed and the preset target speed , calculate the speed error . The error Substitute into the PID control formula:

[0090] ;

[0091] The proportionality coefficient , integral coefficient , differential coefficient Pre-tuned by critical proportionality method, the initial value is set to =0.5, =0.1, =0.2. The control amount is calculated After that, the edge computing gateway sends the control command to the motor driver to adjust the motor speed. For example, when >0 to increase the speed, When the value is less than 0, the speed is reduced to make the centrifuge tube running state gradually stable. During the adjustment process, the speed and other related parameters are continuously monitored, and the PID parameters are dynamically adjusted according to the actual situation to ensure the control effect.

[0092] In some embodiments, parameter inversion and model updating process: When a “model mismatch” is triggered ( ) or the liquid level h changes by more than 5%, the sample density When changes occur, the parameter inversion process is started. Taking the heat dissipation coefficient k as an example, the objective function is set as:

[0093] ;

[0094] in is the measured temperature, To include the heat dissipation coefficient The RBF surface model predicts the temperature. Initialize the heat dissipation coefficient =0.2, learning rate is 0.01, and the objective function is calculated by automatic differentiation technology. Gradient . According to the gradient update formula Perform iterative calculations, and calculate the current objective function value each time iterates ,when (Preset convergence threshold, such as ), it is considered to have reached convergence and the updated heat dissipation coefficient is obtained . The updated Substitute the heat conduction model in S3 and retrain the RBF surface model. Use the latest parameters and data to optimize the model to ensure that the model can accurately reflect the current operating status of the centrifuge tube and complete the entire closed-loop feedback control process.

[0095] The above embodiments are intended to illustrate the present invention, not to limit the present invention. Therefore, changes in illustrative values or substitutions of equivalent components should still fall within the scope of the present invention.

[0096] From the above detailed description, it will be clear to those skilled in the art that the present invention can indeed achieve the aforementioned objectives and is in compliance with the provisions of the Patent Law.

[0097] Although preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic inventive concepts. Therefore, the appended claims are intended to be interpreted as covering the preferred embodiments and all changes and modifications that fall within the scope of the invention. The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

[0098] It should be noted that the above description of the relevant processes is for illustration and purpose only and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the processes under the guidance of this specification. However, such modifications and changes are still within the scope of this specification.

[0099] The basic concepts have been described above. It will be apparent to those skilled in the art after reading this application that the above disclosures are merely illustrative and do not constitute limitations on this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and amendments to this application. Such modifications, improvements, and amendments are suggested in this application and remain within the spirit and scope of the exemplary embodiments of this application.

[0100] At the same time, this application uses specific terms to describe the embodiments of this application. For example, "one embodiment," "an embodiment," and / or "some embodiments" refer to a certain feature, structure, or characteristic related to at least one embodiment of this application. Therefore, it should be emphasized and noted that "one embodiment," "an embodiment," or "an alternative embodiment" mentioned twice or more in different places in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this application may be appropriately combined.

[0101] Furthermore, those skilled in the art will appreciate that various aspects of the present application may be illustrated and described in terms of a number of patentable categories or situations, including any new and useful process, machine, product, or combination of substances, or any new and useful improvement thereof. Thus, various aspects of the present application may be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software. Each of the above hardware and software may be referred to as a "unit," "module," or "system." Furthermore, various aspects of the present application may take the form of a computer program product embodied in one or more computer-readable media, with computer-readable program code embodied therein.

[0102] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, Python, etc., conventional procedural programming languages such as C programming language, Visual Basic, Fortran2103, Perl, COBOL2102, PHP, ABAP, dynamic programming languages such as Python, Ruby and Groovy, or other programming languages. The program code can be run entirely on the user's computer, or as a standalone software package on the user's computer, or partly on the user's computer and partly on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network form, such as a local area network (LAN) or a wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as software as a service (SaaS).

[0103] In addition, unless expressly stated in the claims, the order of the processing elements and sequences described in this application, the use of alphanumeric characters, or the use of other names are not intended to limit the order of the processes and methods of this application. Although the above disclosure discusses some embodiments of the invention that are currently considered useful through various examples, it should be understood that such details are only for illustrative purposes, and the attached claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that are consistent with the essence and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a pure software solution, for example, installation on an existing server or mobile device.

[0104] Similarly, it should be noted that in order to simplify the presentation of this disclosure and thereby facilitate understanding of one or more of the invention's embodiments, the foregoing descriptions of the embodiments of this disclosure sometimes combine multiple features into a single embodiment, figure, or description thereof. However, this approach should not be interpreted as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject matter of the invention may possess fewer features than the single embodiment described above.

Claims

1. A sensor network-based intelligent monitoring method for ultracentrifuge tubes, characterized in that: include: S1. Deploy multiple types of sensors in the centrifugal equipment to form a three-dimensional distributed monitoring network and access the edge computing gateway; the multiple types of sensors include a three-axis accelerometer, a temperature sensor, a pressure sensor, and a liquid level sensor; S2. The edge computing gateway synchronously collects all sensor data and preprocesses it; S3. Develop a centrifugal force-stress model and a heat conduction model based on the centrifuge tube specifications. Calibrate the model parameters using historical data. Fit the bracket vibration transfer function using no-load operation data. Determine the relationship between liquid sloshing and liquid level through load testing. S4. Temporally and spatially align the preprocessed sensor data with the output of the centrifugal force-stress model, and use an extended Kalman filter to fuse multi-source information to estimate the real-time state vector of the centrifuge tube; extract characteristic parameters based on the real-time state vector of the centrifuge tube; the real-time state vector includes speed, temperature, pressure, liquid level, and peak vibration acceleration; the characteristic parameters include speed change rate, temperature rise rate, pressure fluctuation coefficient, vibration energy ratio, liquid level change gradient, and centrifugal force deviation rate; S5. Input the feature parameters into a random forest classifier to identify the state category; If it is determined to be abnormal, a first-level warning, a second-level warning, or a third-level warning will be triggered according to the preset rules; S6. If a level 1 or level 2 warning is detected, the edge computing gateway sends a control instruction to the centrifugal device to adjust the motor speed through the PID controller; at the same time, the parameters of the bracket vibration transfer function, centrifugal force-stress model, or heat conduction model are dynamically updated based on real-time monitoring data.

2. The ultracentrifuge tube intelligent monitoring method based on a sensor network according to claim 1, characterized in that: Multiple types of sensors are deployed, including: a three-axis acceleration sensor attached to the central axis of the centrifuge tube side wall, a temperature sensor embedded in the bottom of the centrifuge tube, a pressure sensor integrated on the inside of the tube cover, and a liquid level sensor fixed at 1 / 2 the height of the tube body using an infrared tube structure.

3. The ultracentrifuge tube intelligent monitoring method based on sensor network according to claim 2, characterized in that: The preprocessing in S2 includes: using a low-pass filter to remove high-frequency noise from the original acceleration signal, using a sliding average filter to eliminate random interference from the original temperature signal, and using threshold verification to eliminate abnormal values from the original pressure and liquid level signals.

4. The method for intelligent monitoring of ultracentrifuge tubes based on a sensor network according to claim 1, wherein: The centrifugal force-stress model is constructed based on the centrifugal force formula, and the heat conduction model is established according to the specific heat capacity and heat dissipation coefficient of the pipe material.

5. The ultracentrifuge tube intelligent monitoring method based on sensor network according to claim 1, characterized in that: The state transfer function of the extended Kalman filter is determined according to the change law of the centrifuge tube speed, and the process noise covariance matrix is obtained through sensor error statistics.

6. The sensor network-based intelligent monitoring method for ultracentrifuge tubes according to claim 1, characterized in that: The training samples of the random forest classifier include normal operating condition data and various abnormal operating condition historical data.

7. The sensor network-based intelligent monitoring method for ultracentrifuge tubes according to claim 1, characterized in that: The first-level warning displays a warning message on the device screen, the second-level warning sends a text message notification to the monitoring terminal and records the abnormality log, and the third-level warning automatically triggers the equipment shutdown program and cuts off the motor power supply.

8. The sensor network-based intelligent monitoring method for ultracentrifuge tubes according to claim 1, characterized in that: The proportional, integral and differential coefficients of the PID controller are adjusted by the critical proportionality method.

9. The sensor network-based intelligent monitoring method for ultracentrifuge tubes according to claim 1, characterized in that: In S6, when the liquid level changes by more than 5%, the total mass of the centrifuge tube is recalculated, including the mass of the sample liquid and the mass of the tube body.

10. The sensor network-based intelligent monitoring method for ultracentrifuge tubes according to claim 1, characterized in that: In S6, by comparing the residuals of the actual monitoring data with the model prediction data, when the residual exceeds a certain threshold, the model parameters are recalibrated.

Images