Instrument full life cycle management method and system based on Internet of Things

Abstract

The invention provides an instrument full life cycle management method and system based on the Internet of Things, and relates to the technical field of the Internet of Things, and the method comprises the steps: binding an RFID tag with an instrument to obtain a unique identifier, collecting operation parameters, and building a multi-dimensional data structure; performing time sequence clustering analysis on the data to identify an abnormal mode, and constructing an equipment group association graph to calculate fault propagation characteristics; constructing a performance attenuation curve through multi-head self-attention feature extraction to generate a prediction maintenance index; optimizing and executing a maintenance task allocation scheme; and updating the prediction index and the distribution result. According to the invention, real-time monitoring and predictive maintenance of the instrument state are realized, and the equipment management efficiency and the operation reliability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, and in particular to an instrument full life cycle management method and system based on Internet of Things. BACKGROUND

[0002] With the development of science and technology, instruments and equipment in laboratories and industrial environments are increasingly complex and intelligent. These instruments play an important role in scientific research, production and detection. Full life cycle management of instruments and equipment is a key link to ensure efficient operation, prolong service life and reduce maintenance costs. Traditional instrument management mainly relies on manual records and regular inspections. With the development of Internet of Things technology, it is possible to collect instrument operation data in real time through sensor networks, and to apply big data analysis and artificial intelligence technology for intelligent management. Currently, Internet of Things technology has been widely used in equipment management, including automatic identification technology such as RFID tags for asset tracking, condition monitoring technology for real-time parameter collection, and data analysis technology for equipment health assessment.

[0003] However, the existing instrument management method based on Internet of Things has some obvious defects. The existing technology usually only focuses on the state monitoring of a single instrument, lacks analysis of the interrelation between instrument groups, and cannot effectively identify and handle the fault propagation characteristics between devices, resulting in insufficient system risk warning capability. Traditional maintenance strategies are mainly based on fixed time intervals or simple threshold triggers, lack precise modeling of instrument performance degradation rules, and cannot accurately predict the optimal maintenance time, often leading to resource waste due to premature maintenance or loss due to late maintenance. Existing technologies mostly use manual experience decision-making in maintenance task allocation, lack adaptive scheduling mechanisms considering multi-dimensional resource constraints and task priority, and are difficult to achieve optimal balance between maintenance timeliness and resource utilization efficiency, limiting the improvement of overall maintenance efficiency. SUMMARY

[0004] The present application provides an instrument full life cycle management method and system based on Internet of Things, which can solve the problems in the prior art.

[0005] In a first aspect of the present application, an instrument full life cycle management method based on Internet of Things is provided, comprising: binding an RFID tag with an instrument to obtain unique identification information, collecting operation parameters of the instrument, and establishing a multi-dimensional data structure; performing time series clustering analysis on the multi-dimensional data structure, identifying abnormal patterns of instrument operation, constructing a device group correlation graph, calculating fault propagation characteristics according to fault state transitions between nodes in the device group correlation graph, and forming a device operation state evaluation result; Based on the equipment operation state evaluation result, a performance attenuation curve of the instrument is constructed through multi-head self-attention feature extraction and fitting function optimization, probability calculation is performed in combination with the fault propagation feature, and a predicted maintenance index is generated; Based on the predicted maintenance index, a scheduling optimization algorithm is adopted to generate a maintenance task allocation scheme with timeliness of maintenance and resource utilization rate as targets; Based on the maintenance task allocation result, maintenance work is performed, maintenance process and result data are stored in a maintenance knowledge base, and the predicted maintenance index and the maintenance task allocation result are updated through online analysis.

[0006] In an optional embodiment, time series clustering analysis is performed on the multi-dimensional data structure, abnormal patterns of instrument operation are identified, and a device group association graph is constructed, including: Time series feature extraction is performed on the multi-dimensional data structure, statistical features, frequency domain features and time-frequency features in the multi-dimensional data structure are extracted, an optimal feature subset is determined according to feature importance weights, and a reduced dimension feature matrix is generated; Based on the reduced dimension feature matrix, time series similarity is calculated using dynamic time warping, a dynamic time warping distance matrix is constructed, and time series clustering results are obtained by iteratively calculating the similarity between samples and clustering centers; The time series clustering results are compared with a preset normal operation mode benchmark, a clustering deviation value is calculated, and abnormal patterns of instrument operation are generated according to a comparison result of the clustering deviation value and a preset abnormality determination threshold; Based on the abnormal patterns of instrument operation, a device group association graph is constructed, a correlation coefficient between devices is calculated based on device operation state parameters, a connection relationship between nodes is determined based on an association strength threshold, an edge weight is calculated based on device physical locations, operation parameters and the abnormal patterns, and a weighted device group association graph is generated.

[0007] In an optional embodiment, the fault propagation feature is calculated based on the fault state transition between nodes in the device group association graph, and the equipment operation state evaluation result is formed, including: The number of fault state transitions between adjacent nodes in the device group association graph is counted in a preset time window, the number of fault state transitions is divided by the length of the time window to obtain a fault transition probability between nodes, and a fault propagation probability matrix is formed; The shortest propagation path and propagation time delay between each pair of nodes in the fault propagation probability matrix are calculated by gradually updating the propagation distance between nodes, nodes with a fault transition probability greater than a preset probability threshold are marked as key propagation nodes, and a fault propagation path graph is generated; The number of connecting edges of each node in the fault propagation path graph is counted. The number of connecting edges is multiplied by the fault propagation probability of each edge to calculate the node influence degree. The ratio of the fault propagation distance between nodes to the sampling interval is calculated to obtain the fault spread rate. Areas where the node influence degree and the fault spread rate exceed the corresponding threshold are marked as propagation danger zones. The distance from the edge nodes of the propagation danger zone to the key propagation nodes is calculated to determine the fault impact range. The mean and variance of equipment operating parameters are calculated to obtain the single equipment status index, the correlation coefficient between parameters of equipment is calculated to obtain the group coordination index, the risk index is determined by weighting the node influence degree, diffusion speed and fault impact range, and the weighted average of the single equipment status index, group coordination index and risk index is used to obtain the equipment operating status assessment result.

[0008] In an optional embodiment, based on the device operating status assessment results, constructing the instrument's performance degradation curve through multi-head self-attention feature extraction and fitting function optimization includes: The device operation status evaluation results are input into the multi-head self-attention layer, which includes a query-key-value calculation unit. The query-key-value calculation unit captures features at different time scales and determines the attention features. The attention features are input into the temporal position coding layer, and the relative position of the attention features is represented by sine and cosine functions to determine the temporal dependency and generate position coding features. The location encoding features are transformed by dimensional mapping to generate intermediate features. The intermediate features are then passed through an activation function to obtain transformed features. The transformed features are then superimposed with the location encoding features to establish short-circuit connection paths that preserve the original feature information, resulting in fused features. Based on the fusion features, a state change rate sequence is constructed, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features. Based on the time interval and amplitude difference of the state change points, the attenuation feature groups are grouped to construct attenuation feature groups, which are then stored in the attenuation feature database, and the optimal fitting function is selected. Calculate the mean squared error between the multi-scale performance characteristics and the optimal fitting function, construct a gradient descent matrix, adjust the fitting coefficients of the optimal fitting function according to the gradient descent matrix, repeat the iteration until the mean squared error is less than a preset termination threshold, and obtain the instrument performance degradation curve.

[0009] In one optional embodiment, a state change rate sequence is constructed based on the fused features, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features, including: The difference values ​​between adjacent time points are calculated for the fused features according to the time series to construct a state change rate sequence. The difference values ​​are then calculated for the state change rate sequence to construct a performance change acceleration sequence. The state change point is determined based on the local maxima and local minima of the state change rate sequence and the zero point of the performance change acceleration sequence. The fluctuation patterns of the state change points within different time windows are extracted to construct a state fluctuation feature sequence; periodic change features are extracted based on the recurrence frequency and amplitude changes of the state fluctuation feature sequence; and a moving average is applied to the state change points to extract the baseline trend of state changes as a long-term evolution feature. By weighting and combining periodic change characteristics and long-term evolution characteristics, multi-scale performance characteristics are generated.

[0010] In one optional embodiment, the scheduling optimization algorithm includes: A two-layer coding structure for maintenance tasks is constructed based on predictive maintenance indicators. The two-layer coding structure includes an outer path coding for determining the execution order of maintenance tasks and an inner time coding for determining the execution time of maintenance tasks. In the outer path encoding, the initial pheromone distribution between maintenance tasks is set, and the transition probability matrix between maintenance tasks is constructed based on the maintenance timeliness index; A time allocation space for maintenance tasks is established in the inner time encoding, and a time allocation probability matrix for maintenance tasks is constructed based on the resource utilization index. The path search for the execution order of maintenance tasks is performed based on the transition probability matrix, and the execution time of maintenance tasks is determined based on the time allocation probability matrix. Using the time window constraint and resource constraint in the predicted maintenance index as temperature parameters, the maintenance task allocation scheme obtained by the search of the dual-layer coding structure is locally adjusted, and the initial pheromone distribution is updated based on the adjusted objective function value. Based on the two-layer coding structure, the scheduling optimization algorithm is repeatedly executed until the preset number of iterations is reached, and a maintenance task allocation scheme is output.

[0011] In an optional embodiment, the path search for the execution order of maintenance tasks is performed based on the transition probability matrix, and the execution time of the maintenance tasks is determined based on the time allocation probability matrix, further comprising: In the outer path encoding, the execution order position of the next maintenance task is selected based on the transition probability matrix; In the inner time encoding, the execution time of the maintenance task corresponding to the execution order position is determined based on the time allocation probability matrix; The execution result of the maintenance task at the specified execution time is fed back to the outer path encoding to update the transition probability between the maintenance task and other maintenance tasks in the transition probability matrix.

[0012] A second aspect of this invention provides an Internet of Things (IoT)-based instrument lifecycle management system, comprising: The first unit is used to bind RFID tags to instruments to obtain unique identification information, collect instrument operating parameters, and establish a multi-dimensional data structure. The second unit is used to perform time-series clustering analysis on the multidimensional data structure, identify abnormal operating patterns of the instrument, construct an equipment group association diagram, calculate fault propagation characteristics based on the fault state transfer between nodes in the equipment group association diagram, and form an equipment operating status assessment result. The third unit is used to construct the instrument's performance degradation curve based on the equipment operation status assessment results by extracting multi-head self-attention features and optimizing fitting functions, and to perform probability calculations by combining the fault propagation features to generate predictive maintenance indicators. The fourth unit is used to generate a maintenance task allocation scheme based on the predicted maintenance indicators and using a scheduling optimization algorithm with the goals of maintenance timeliness and resource utilization. The fifth unit is used to perform maintenance work based on the maintenance task allocation results, store the maintenance process and result data in the maintenance knowledge base, and update the predicted maintenance indicators and the maintenance task allocation results through online analysis.

[0013] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0015] In this embodiment of the invention, RFID tags are bound to instruments using Internet of Things (IoT) technology to obtain unique identifiers. Through multi-dimensional data structures and time-series clustering analysis, real-time monitoring of instrument operating status and accurate identification of abnormal patterns are achieved, providing accurate fault warnings and diagnostic basis for equipment maintenance and significantly reducing equipment failure rates. Based on multi-head self-attention feature extraction and performance degradation curves constructed using fitting function optimization, combined with probability calculations of fault propagation characteristics, scientific prediction of instrument maintenance needs is achieved, improving the accuracy of maintenance decisions, reducing unnecessary preventative maintenance, lowering maintenance costs, and extending equipment lifespan. The maintenance task allocation scheme generated through scheduling optimization algorithms balances maintenance timeliness and resource utilization, achieving reasonable allocation of maintenance resources. Simultaneously, the establishment of a maintenance knowledge base and an online analysis and update mechanism form a closed-loop feedback system, continuously optimizing predicted maintenance indicators and task allocation schemes, thereby improving overall equipment management efficiency and reliability. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the IoT-based instrument lifecycle management method according to an embodiment of the present invention. Figure 2 Flowchart of the algorithm for optimizing predictive maintenance task scheduling. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0019] Figure 1 This is a flowchart illustrating the IoT-based instrument lifecycle management method according to an embodiment of the present invention. Figure 1 As shown, the method includes: Bind RFID tags to instruments to obtain unique identification information, collect instrument operating parameters, and establish a multi-dimensional data structure; Temporal clustering analysis is performed on the multidimensional data structure to identify abnormal operating patterns of the instrument, and an equipment group association graph is constructed. Based on the fault state transition between nodes in the equipment group association graph, fault propagation characteristics are calculated to form equipment operating status assessment results. Based on the equipment operation status assessment results, the performance degradation curve of the instrument is constructed by extracting multi-head self-attention features and optimizing fitting functions. Probability calculation is performed in combination with the fault propagation features to generate predictive maintenance indicators. Based on the predicted maintenance indicators, a scheduling optimization algorithm is used to generate a maintenance task allocation scheme with the goals of maintenance timeliness and resource utilization. Maintenance work is performed based on the maintenance task allocation results, and the maintenance process and result data are stored in the maintenance knowledge base. The predicted maintenance indicators and the maintenance task allocation results are updated through online analysis.

[0020] In one optional implementation, performing time-series clustering analysis on the multidimensional data structure to identify abnormal patterns in instrument operation and constructing a device group association graph includes: Temporal feature extraction is performed on multidimensional data structures, extracting statistical features, frequency domain features, and time-frequency features from the multidimensional data structures. The optimal feature subset is determined based on the feature importance weights, and a dimension-reduced feature matrix is ​​generated. Based on the dimensionality-reduced feature matrix, dynamic time warping is used to calculate the similarity of time series sequences, a dynamic time warping distance matrix is ​​constructed, and the time series clustering results are obtained by iteratively calculating the similarity between samples and cluster centers; The time-series clustering results are compared with a preset normal operation mode benchmark to calculate the clustering deviation value. Based on the comparison result of the clustering deviation value and the preset anomaly judgment threshold, an abnormal mode of instrument operation is generated. Based on the abnormal operation patterns of the instruments, a device group association graph is constructed. The device operation status parameters are extracted to calculate the correlation coefficient between devices, the association strength threshold is determined to establish node connection relationships, and the edge weights are calculated based on the device physical location, operation parameters and the abnormal patterns to generate a weighted device group association graph.

[0021] In one specific implementation, when performing feature extraction and anomaly identification on multidimensional time-series data of IoT instruments and devices, the first step is to acquire multidimensional time-series data of the instruments, including multiple parameters such as temperature, pressure, flow rate, voltage, and current during instrument operation, forming a multidimensional data structure. Time-series feature extraction is then performed on the acquired multidimensional data structure, including the calculation of statistical features, frequency domain features, and time-frequency features.

[0022] During statistical feature extraction, the mean, standard deviation, kurtosis, skewness, and interquartile range (IVR) of each dimension of data are calculated. Taking temperature data as an example, within a typical 24-hour monitoring cycle, the average temperature of a normally operating instrument is 37.5 degrees Celsius, with a standard deviation of 1.2, kurtosis of 2.8, skewness of 0.3, and IVR of 2.1 degrees Celsius. During frequency domain feature extraction, Fourier transform is performed on the time-series data to extract features such as power spectral density, dominant frequency components and their amplitudes, and frequency band energy distribution. For example, after transforming voltage fluctuation data, a characteristic distribution can be obtained where the energy proportion in the low-frequency band (0-10 Hz) is 65%, the energy proportion in the mid-frequency band (10-50 Hz) is 30%, and the energy proportion in the high-frequency band (above 50 Hz) is 5%. During time-frequency feature extraction, wavelet transform is used to capture the changing characteristics of the data at different time scales, and wavelet energy and wavelet entropy are calculated.

[0023] After feature extraction, the importance of all extracted features is evaluated using a tree-based feature importance calculation method, assigning a weight value to each feature. Based on actual testing, the standard deviation of flow rate fluctuation has a weight of 0.85, the dominant frequency component of voltage has a weight of 0.76, and the wavelet entropy of temperature has a weight of 0.73. These features constitute a subset of features with high importance. Based on a set feature importance threshold (e.g., 0.5), features with importance higher than the threshold are selected to form the optimal feature subset, generating the dimensionality-reduced feature matrix.

[0024] Based on the dimensionality-reduced feature matrix, a dynamic time warping algorithm is used to calculate the similarity between time series sequences. For any two time series sequences, the dynamic time warping distance is calculated, which measures the alignment cost of the best match between the two sequences on the time axis. During calculation, a cumulative distance matrix is ​​constructed, and the optimal path is found through dynamic programming, with the minimum matching cost being used as the distance value between the two sequences. The dynamic time warping distance is calculated pairwise for all samples, and a dynamic time warping distance matrix is ​​constructed.

[0025] Temporal clustering analysis is performed based on the constructed distance matrix. A partitioning clustering method is adopted, initially randomly selecting cluster centers, iteratively calculating the similarity between samples and cluster centers, assigning samples to the most similar clusters, updating cluster centers, and continuing until the clustering results stabilize.

[0026] The obtained time-series clustering results are compared with a preset normal operating mode benchmark. The normal operating mode benchmark is a standard operating mode obtained through historical data statistics, including the standard location and distribution range of each cluster center. The clustering deviation value between the current clustering result and the benchmark mode is calculated. The deviation value includes two aspects: the shift in the location of the cluster center and the change in the cluster distribution pattern. For example, taking voltage parameters as an example, the cluster center value in the normal operating mode is 220 volts, and the cluster radius is 5 volts; while the current cluster result has a center value of 228 volts and a cluster radius of 8 volts, and the calculated deviation value is 13.6.

[0027] The calculated clustering deviation value is compared with a preset anomaly detection threshold, which is set based on historical anomaly cases and expert experience, with a typical value of 10.0. If the deviation value exceeds the threshold, it is determined to be an abnormal state, and the anomaly type is determined based on the deviation characteristics. In the above case, the voltage parameter deviation value of 13.6 is greater than the threshold of 10.0, which is determined to be a voltage anomaly, and an anomaly mode identifier for instrument operation is generated.

[0028] Based on the detected abnormal instrument operation patterns, a device group correlation diagram is constructed. Device operating status parameters are extracted, including the operating parameters and abnormality markers of each device. Correlation coefficients between devices are calculated using the Pearson correlation coefficient method to measure the correlation of device operating parameters. For example, the measured temperature correlation coefficient between two adjacent temperature control devices is 0.86, indicating a high correlation between their operating states.

[0029] Based on the calculated correlation coefficient, a threshold for association strength is determined, with a typical value of 0.7. When the correlation coefficient between devices exceeds the threshold, a node connection is established. Edge weights are calculated based on the physical location information of the devices, the similarity of operating parameters, and the consistency of anomaly patterns. The edge weight calculation comprehensively considers physical distance (weight 0.3), parameter correlation (weight 0.5), and anomaly pattern consistency (weight 0.2). For example, two instruments with the same voltage anomaly pattern, a physical distance of 10 meters, a parameter correlation of 0.85, and an anomaly pattern consistency of 1.0, would have a calculated edge weight of 0.805.

[0030] The above calculations generate a weighted device group association graph. Nodes represent individual devices, edges represent the relationships between devices, and the weight of the edges reflects the strength of the relationship. Edges with high weights (e.g., weight greater than 0.8) are marked as strong associations; edges with moderate weights (e.g., weight between 0.7 and 0.8) are marked as medium associations; and edges with low weights (e.g., weight less than 0.7) are not displayed in the graph. The final association graph visually displays the anomaly propagation path and key nodes in the device group, providing a basis for anomaly tracing and fault diagnosis.

[0031] In one optional implementation, the fault propagation characteristics are calculated based on the fault state transitions between nodes within the equipment group association diagram to form the equipment operating status assessment results, including: Within a preset time window, the number of fault state transitions between adjacent nodes in the device group association diagram is counted. The number of fault state transitions is divided by the length of the time window to obtain the fault transition probability between nodes, forming a fault propagation probability matrix. By gradually updating the propagation distance between nodes, the shortest propagation path and propagation time delay between each pair of nodes in the fault propagation probability matrix are calculated. Nodes with a fault transfer probability greater than a preset probability threshold are marked as key propagation nodes, and a fault propagation path map is generated. The number of connecting edges of each node in the fault propagation path graph is counted. The number of connecting edges is multiplied by the fault propagation probability of each edge to calculate the node influence degree. The ratio of the fault propagation distance between nodes to the sampling interval is calculated to obtain the fault spread rate. Areas where the node influence degree and the fault spread rate exceed the corresponding threshold are marked as propagation danger zones. The distance from the edge nodes of the propagation danger zone to the key propagation nodes is calculated to determine the fault impact range. The mean and variance of equipment operating parameters are calculated to obtain the single equipment status index, the correlation coefficient between parameters of equipment is calculated to obtain the group coordination index, the risk index is determined by weighting the node influence degree, diffusion speed and fault impact range, and the weighted average of the single equipment status index, group coordination index and risk index is used to obtain the equipment operating status assessment result.

[0032] In one specific implementation, device operation data within a preset time window is acquired, for example, a 30-day time window. During this period, the operating status of each node in the device group association graph is continuously monitored. Assuming there are five device nodes A, B, C, D, and E in the association graph, the fault occurrence and timestamp of each node are recorded. When a fault is detected in node A, if node B also fails within a short period (e.g., within one hour), a fault state transition from A to B is recorded. Within the 30-day time window, a total of 15 fault state transitions occur from A to B, so the fault transition probability from A to B is 15 / 30 = 0.5. Similarly, the fault transition probabilities between all adjacent node pairs are calculated to form a fault propagation probability matrix. For example, this matrix might show: the probability from node A to B is 0.5, the probability from B to C is 0.4, the probability from C to D is 0.6, the probability from D to E is 0.3, and the probability from A to C is 0.1, etc.

[0033] Based on the aforementioned fault propagation probability matrix, the shortest propagation path and propagation time delay between each node pair are calculated by iteratively updating the propagation distance between nodes. Specifically, an improved shortest path algorithm is used, initializing the propagation distance between all node pairs to infinity, and then finding the shortest path through iterative updates. For example, the shortest path from node A to node E might be A→B→C→D→E, with a corresponding propagation probability of 0.5×0.4×0.6×0.3=0.036, and the propagation time delay being the sum of the propagation times of each segment, assumed to be 4 hours. In practical applications, a preset probability threshold of 0.4 can be set, and nodes with a fault transfer probability greater than 0.4 are marked as critical propagation nodes. In this embodiment, nodes B, C, and D are marked as critical propagation nodes. By connecting these critical nodes and their propagation paths, a fault propagation path graph is generated.

[0034] The number of connecting edges for each node in the fault propagation path graph is counted. For example, node C is connected to an edge from B and an edge leading to D, for a total of 2 connecting edges. The influence of node C is calculated as the number of connecting edges multiplied by the average fault propagation probability of each edge, i.e., 2 × (0.4 + 0.6) / 2 = 1.0. Simultaneously, the ratio of the fault propagation distance between nodes to the sampling interval is calculated to obtain the fault propagation speed. Assuming the physical distance from B to C is 100 meters, the fault propagation time is 30 minutes, and the system sampling interval is 5 minutes, then the fault propagation speed is (100 / 30) × 5 = 16.67 meters / minute. Areas with a node influence exceeding 1.0 and a fault propagation speed exceeding 15 meters / minute are marked as propagation danger zones. In this embodiment, the area where node C is located is marked as a propagation danger zone. By calculating the distance from the edge nodes of the propagation danger zone (such as nodes B and D) to the critical propagation node (node ​​C), the fault influence range is determined to be 100 meters.

[0035] To comprehensively assess the operational status of equipment, individual equipment status indicators, group coordination indicators, and risk indicators are calculated. Individual equipment status indicators are obtained by calculating the mean and variance of the equipment's operating parameters. For example, for node C, its temperature sensor has a mean of 60℃ and a variance of 2.5 during normal operation, while the current measured mean is 65℃ and the variance is 4.0, indicating that the equipment may be in an abnormal state. Group coordination indicators are obtained by calculating the correlation coefficient between parameters of different devices. For example, the correlation coefficient between the temperature parameters of nodes B and C is 0.85, indicating a strong coordination relationship between the two devices. Risk indicators are determined based on a weighted average of node influence, diffusion rate, and fault impact range. Assuming the weights for the three are 0.4, 0.3, and 0.3 respectively, the risk indicator for node C is calculated as: 0.4 × 1.0 + 0.3 × (16.67 / 15) + 0.3 × (100 / 100) = 0.4 + 0.333 + 0.3 = 1.033.

[0036] The weighted average of individual device status indicators, group coordination indicators, and risk indicators yields the equipment operation status assessment result. The weights for the three indicators are set to 0.3, 0.3, and 0.4, respectively. For node C, assuming its individual device status indicator is 0.8 (considering temperature anomalies), its group coordination indicator is 0.85, and its risk indicator is 1.033, then its final operation status assessment result is 0.3×0.8 + 0.3×0.85 + 0.4×1.033 = 0.24 + 0.255 + 0.4132 = 0.9082. This value is close to 1, indicating that device C has a high operation risk and requires close monitoring.

[0037] In this embodiment, by analyzing the fault propagation characteristics in the device group association diagram and comprehensively considering the status of individual devices, group coordination, and risk factors, a comprehensive and objective basis for assessing the operating status of devices is provided, which helps to prevent fault propagation and systemic risks, and improves the overall reliability and security of the device group.

[0038] In one optional implementation, based on the device operating status assessment results, constructing the instrument's performance degradation curve through multi-head self-attention feature extraction and fitting function optimization includes: The device operation status evaluation results are input into the multi-head self-attention layer, which includes a query-key-value calculation unit. The query-key-value calculation unit captures features at different time scales and determines the attention features. The attention features are input into the temporal position coding layer, and the relative position of the attention features is represented by sine and cosine functions to determine the temporal dependency and generate position coding features. The location encoding features are transformed by dimensional mapping to generate intermediate features. The intermediate features are then passed through an activation function to obtain transformed features. The transformed features are then superimposed with the location encoding features to establish short-circuit connection paths that preserve the original feature information, resulting in fused features. Based on the fusion features, a state change rate sequence is constructed, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features. Based on the time interval and amplitude difference of the state change points, the attenuation feature groups are grouped to construct attenuation feature groups, which are then stored in the attenuation feature database, and the optimal fitting function is selected. Calculate the mean squared error between the multi-scale performance characteristics and the optimal fitting function, construct a gradient descent matrix, adjust the fitting coefficients of the optimal fitting function according to the gradient descent matrix, repeat the iteration until the mean squared error is less than a preset termination threshold, and obtain the instrument performance degradation curve.

[0039] In one specific implementation, during the instrument's full lifecycle management process, after acquiring the equipment operating status assessment results, they are input into a multi-head self-attention layer for processing. This multi-head self-attention layer contains eight parallel query-key-value calculation units, each with a 64-dimensional feature space, generating query vectors, key vectors, and value vectors through linear transformations. Specifically, the equipment operating status assessment results are represented as a 512-dimensional vector sequence, with each vector containing assessment values ​​for parameters such as temperature, humidity, voltage, and current. For the temperature parameter, the normal operating value is 36.5, the assessment result is 35.8, and the difference is 0.7; for the voltage parameter, the normal value is 220, the assessment result is 218, and the difference is 2. These difference values ​​constitute a part of the assessment result vector.

[0040] In the multi-head self-attention processing, each query-key-value computation unit is responsible for capturing features at different temporal scales. The first computation unit focuses on short-term temporal features with an attention window of 1 hour; the second focuses on medium-term temporal features with an attention window of 6 hours; and the third focuses on long-term temporal features with an attention window of 24 hours. An attention score is obtained by calculating the dot product of the query vector and the key vector and performing scaling. For the temperature parameter, the attention score is 0.82 in the short-term window, 0.65 in the medium-term window, and 0.43 in the long-term window, indicating that temperature changes are more correlated in the short term. The attention score is applied to the value vector to obtain a weighted feature representation. The output features from the eight heads are merged and a linear transformation is performed to obtain the final attention feature with a dimension of 512.

[0041] Attention features are input into a temporal positional encoding layer, and sine and cosine functions are used to represent the relative positions of these features. Positional encoding uses sine and cosine functions of different frequencies to generate a unique encoded value for each position in the sequence. In the 512-dimensional feature space, sine functions are used for even-numbered dimensions, and cosine functions are used for odd-numbered dimensions. The period of the positional encoding varies from 2π to 10000×2π, covering various time scales from short-term to long-term. For example, for the 10th position in the sequence, the positional encoding value is 0.841 in dimension 0, 0.540 in dimension 1, 0.005 in dimension 510, and 0.999 in dimension 511. By adding the positional encoding to the attention features, temporal dependencies are determined, generating positional encoded features.

[0042] The location-encoded features undergo a dimensionality mapping transformation. A fully connected layer maps the 512-dimensional features to a 2048-dimensional space, generating intermediate features. A parameterized rectified linear unit (RCU) is used as the activation function to perform a non-linear transformation on the intermediate features. This activation function maintains a linear relationship for positive inputs and scales the values ​​to small values ​​close to but not zero for negative inputs, with a slope parameter set to 0.1. After activation function processing, the transformed features are obtained, still maintaining a 2048-dimensional dimension. These transformed features are then mapped back to the 512-dimensional space through another fully connected layer and element-wise added to the location-encoded features, establishing short-circuit connections to preserve the original feature information and obtain fused features. Short-circuit connections allow gradients to flow directly back to earlier layers, mitigating the vanishing gradient problem and improving model training stability.

[0043] A state change rate sequence is constructed based on fused features, and the change rate is obtained by calculating the difference between the fused features at adjacent time points. For the temperature parameter, the average change rate sequence over 24 hours under normal conditions is 0.1 / hour, with a standard deviation of 0.05; while under abnormal conditions, the average change rate rises to 0.3 / hour, and the standard deviation increases to 0.15. State change points are determined by setting a change rate threshold (e.g., 0.25 / hour). Periodic change features of the state change points are extracted, including change frequency, amplitude, and phase. For example, the temperature parameter shows a peak every 24 hours under normal operating conditions, with a peak deviation not exceeding 0.5. Long-term evolution features are also extracted, including the trend line slope and the long-term modulation effect of periodic changes. For example, the voltage parameter shows a decreasing trend of -0.02 / day during equipment aging. Multi-scale performance features are generated by combining periodic change features and long-term evolution features.

[0044] The data are grouped based on the temporal intervals and amplitude differences of state change points. The temporal intervals are categorized as short-term (1-24 hours), medium-term (1-7 days), and long-term (over 7 days); the amplitude differences are categorized as minute (change rate <0.1 / hour), moderate (change rate 0.1-0.3 / hour), and significant (change rate >0.3 / hour). Nine feature grouping types are obtained by combining these, constructing attenuation feature groups. For example, for temperature parameters, the short-term-significant group contains 5 change points, the medium-term-moderate group contains 12 change points, and the long-term-minute group contains 8 change points. The attenuation feature groups are stored in an attenuation feature database containing historical attenuation feature grouping data for different instrument types and operating environments.

[0045] Historical data matching the current instrument type and operating environment were retrieved from the attenuation feature database. The optimal fitting function was selected by comparing the similarity between the current attenuation feature group and the historical data. Candidate fitting functions included linear, exponential, power, and logarithmic functions. For motor-type instruments, the exponential function typically provides the best fit; for sensor-type instruments, the power function typically provides the best fit; for electronic components, the linear function provides a good fit in the initial stage, while the exponential function is better in the later stages. Based on historical data statistics, the exponential function was selected as the optimal fitting function for the current instrument, with initial parameters set to a starting value of 1.0 and an attenuation rate of 0.005 / day.

[0046] The mean squared error (MSE) between the multi-scale performance characteristics and the optimal fitting function is calculated as an evaluation metric for the fitting quality. The initial MSE after fitting is 0.285, exceeding the preset termination threshold of 0.05. A gradient descent matrix is ​​constructed, containing the partial derivatives of the MSE with respect to each fitted parameter. The fitting coefficients of the optimal fitting function are adjusted based on the gradient descent matrix, with the learning rate set to 0.01. After the first iteration, the initial value is adjusted to 0.98, the decay rate is adjusted to 0.006 / day, and the MSE decreases to 0.215. After the fifth iteration, the initial value is 0.95, the decay rate is 0.0075 / day, and the MSE decreases to 0.108. After the tenth iteration, the initial value is 0.93, the decay rate is 0.0082 / day, and the MSE decreases to 0.063. After the fifteenth iteration, the initial value is 0.92, the decay rate is 0.0085 / day, and the MSE decreases to 0.048, which is less than the preset termination threshold of 0.05, and the iteration terminates.

[0047] The obtained instrument performance degradation curve is represented as an exponential function, with an initial value of 0.92 and a degradation rate of 0.0085 / day. This curve indicates that the instrument performance will decrease to 50% of its initial performance on day 82 and to 25% on day 165. Based on the degradation curve prediction, the instrument will require maintenance or replacement when its performance drops to 30% (approximately day 140), and a warning should be issued 30 days in advance (i.e., day 110). Accurate modeling of the performance degradation curve enables scientific management of the instrument's entire lifecycle, optimizes maintenance strategies, extends instrument lifespan, and improves operational efficiency.

[0048] In one optional implementation, a state change rate sequence is constructed based on the fused features, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features, including: The difference values ​​between adjacent time points are calculated for the fused features according to the time series to construct a state change rate sequence. The difference values ​​are then calculated for the state change rate sequence to construct a performance change acceleration sequence. The state change point is determined based on the local maxima and local minima of the state change rate sequence and the zero point of the performance change acceleration sequence. The fluctuation patterns of the state change points within different time windows are extracted to construct a state fluctuation feature sequence; periodic change features are extracted based on the recurrence frequency and amplitude changes of the state fluctuation feature sequence; and a moving average is applied to the state change points to extract the baseline trend of state changes as a long-term evolution feature. By weighting and combining periodic change characteristics and long-term evolution characteristics, multi-scale performance characteristics are generated.

[0049] In one specific implementation, the difference between adjacent time points is calculated for the fused feature according to the time series to construct a state change rate sequence. Specifically, assuming the fused feature sequence is F(t), where t represents a time point, t=1, 2, ..., T, and T is the total time length. The difference between adjacent time points is calculated as: ΔF(t)=F(t+1)-F(t), where t=1, 2, ..., T-1. The resulting ΔF(t) is the state change rate sequence, reflecting the rate of change of the system state at each time point. For example, for the equipment temperature fused feature [35.2, 35.4, 35.8, 36.1, 36.5, 36.7, 36.8, 36.7, 36.5, 36.2]℃, the calculated state change rate sequence is [0.2, 0.4, 0.3, 0.4, 0.2, 0.1, -0.1, -0.2, -0.3]℃ / time unit.

[0050] Calculate the difference values ​​for the rate of change of state to construct a performance change acceleration sequence. Specifically, calculate the difference value of ΔF(t): Δ 2 F(t) = ΔF(t+1) - ΔF(t), where t = 1, 2, ..., T-2. The resulting ΔF(t)... 2 F(t) is the performance change acceleration sequence, reflecting the rate of change of the system state. Continuing the example above, the calculated performance change acceleration sequence is [0.2, -0.1, 0.1, -0.2, -0.1, -0.2, -0.1, -0.1, -0.1]℃ / time unit. 2 .

[0051] Based on the local maxima and local minima of the rate of change sequence and the zeros of the acceleration sequence of performance change, the state change points are determined. A local maximum represents the moment when the rate of increase in the state reaches its peak, a local minimum represents the moment when the rate of decrease in the state reaches its peak, and the zeros of the acceleration sequence represent the inflection points where the rate of change in the state changes from increasing to decreasing or from decreasing to increasing. Specifically, traversing the rate of change sequence ΔF(t), for any time t, if ΔF(t) > ΔF(t-1) and ΔF(t) > ΔF(t+1), then time t is a local maximum point; if ΔF(t) < ΔF(t-1) and ΔF(t) < ΔF(t+1), then time t is a local minimum point. For the acceleration sequence of performance change Δ²F(t), if ΔF(t) > ΔF(t-1) and ΔF(t+1), then ΔF(t) > ΔF(t+1), then ΔF(t) > ΔF(t+1), then ΔF(t) > ΔF(t+1). 2 F(t) and Δ 2 If F(t+1) has opposite signs, then there is a zero point between t and t+1. By combining these three types of characteristic points, the key moments of change in the system state can be identified. In the example above, the local maxima are located at time points 3 (corresponding to a rate of change of 0.4) and 5 (corresponding to a rate of change of 0.4), the local minima are located at time point 9 (corresponding to a rate of change of -0.3), and the zero points of the acceleration sequence are located between time points 2 and 4, and between time points 6 and 7.

[0052] The fluctuation patterns of state change points within different time windows are extracted to construct a state fluctuation characteristic sequence. Multiple time windows of different sizes are set, such as 24 hours, 7 days, and 30 days, and the distribution and characteristics of state change points within each window are analyzed. For each window w, the number of state change points n_w within the window is counted, and the average time interval μ_w and standard deviation σ_w between change points, as well as the average amplitude a_w and standard deviation s_w, are calculated. These statistics constitute the state fluctuation characteristic sequence W(w) = [n_w, μ_w, σ_w, a_w, s_w]. For example, within a 24-hour window, the number of temperature change points may be 12, with an average interval of 2 hours, a standard deviation of 0.5 hours, an average amplitude of 0.3℃, and an amplitude standard deviation of 0.1℃.

[0053] Periodic variation features are extracted based on the recurrence frequency and amplitude changes of the state fluctuation feature sequences. Using a sliding window method, for each window of size w, the similarity S(i, j) of the state fluctuation feature sequences of adjacent windows is calculated, where i and j are window indices. When the similarity is higher than a preset threshold θ (e.g., 0.85), the two windows are considered to have similar fluctuation patterns. The periodic pattern of the system is identified by detecting the time interval between similar windows. For each detected period p, its period length T_p, occurrence frequency f_p, and stability index s_p (number of consecutive periods) are recorded. These features constitute the periodic variation feature P = [{T_p, f_p, s_p}]. For example, a 24-hour temperature fluctuation cycle might be detected, with a frequency of 1 time / day and a stability index of 30 (consecutive occurrences for 30 days).

[0054] A moving average is applied to the state change points to extract the baseline trend of the state changes as a long-term evolution feature. A large window size W (e.g., 30 days) is selected, and a moving average is applied to the amplitude values ​​of the state change points to obtain a smoothed trend sequence T(t). The slope change of the trend sequence is calculated to identify key trend turning points. For each trend segment k, its start time t_start_k, end time t_end_k, duration d_k, average slope s_k, and trend direction dir_k (upward or downward) are recorded. These features constitute the long-term evolution feature L=[{t_start_k, t_end_k, d_k, s_k, dir_k}]. For example, an upward trend lasting 45 days with an average slope of 0.02℃ / day might be identified, starting on January 1st and ending on February 15th.

[0055] Multi-scale performance characteristics are generated by weighting and combining periodic variation features and long-term evolution features. Based on the application scenario requirements, the periodic feature weight w_p and the long-term feature weight w_l are set, satisfying w_p + w_l = 1. For each time point t, the periodic feature value F_p(t) and the long-term evolution feature value F_l(t) are calculated, and the final multi-scale performance feature M(t) = w_p·F_p(t) + w_l·F_l(t) is obtained by weighted summation. For example, for applications predicting equipment failures, w_p = 0.4 and w_l = 0.6 might be set, as long-term evolution features may be more important for failure prediction. The generated multi-scale performance features can simultaneously reflect the system's periodic variation patterns and long-term evolution trends, providing effective feature inputs for subsequent state assessment, fault diagnosis, and predictive maintenance.

[0056] In one optional implementation, the scheduling optimization algorithm includes: A two-layer coding structure for maintenance tasks is constructed based on predictive maintenance indicators. The two-layer coding structure includes an outer path coding for determining the execution order of maintenance tasks and an inner time coding for determining the execution time of maintenance tasks. In the outer path encoding, the initial pheromone distribution between maintenance tasks is set, and the transition probability matrix between maintenance tasks is constructed based on the maintenance timeliness index; A time allocation space for maintenance tasks is established in the inner time encoding, and a time allocation probability matrix for maintenance tasks is constructed based on the resource utilization index. The path search for the execution order of maintenance tasks is performed based on the transition probability matrix, and the execution time of maintenance tasks is determined based on the time allocation probability matrix. Using the time window constraint and resource constraint in the predicted maintenance index as temperature parameters, the maintenance task allocation scheme obtained by the search of the dual-layer coding structure is locally adjusted, and the initial pheromone distribution is updated based on the adjusted objective function value. Based on the two-layer coding structure, the scheduling optimization algorithm is repeatedly executed until the preset number of iterations is reached, and a maintenance task allocation scheme is output.

[0057] In one specific implementation, the outer path encoding sets the initial pheromone distribution among maintenance tasks to characterize the execution priority relationship between different maintenance tasks. Taking IoT-connected laboratory analytical instruments as an example, assume there are 5 instruments requiring predictive maintenance, labeled as Task 1 to Task 5. The initial pheromone distribution can be set to a uniform distribution, i.e., the initial pheromone concentration is 0.2 for all tasks. A transition probability matrix between maintenance tasks is constructed based on the maintenance timeliness index obtained from instrument status monitoring. This index reflects the failure risk level of each instrument. For example, based on historical data analysis, the transition probability from Task 1 to Task 2 is 0.35, the transition probability from Task 2 to Task 3 is 0.28, and so on, forming a complete 5×5 transition probability matrix.

[0058] The inner time coding establishes a time allocation space for maintenance tasks, which consists of available time points within the current scheduling cycle. Taking an 8-hour workday as an example, the time allocation space can be divided into 48 15-minute time segments. A time allocation probability matrix for maintenance tasks is constructed based on resource utilization indicators, reflecting the efficiency of executing maintenance tasks at different time points. For each maintenance task, its allocation probability in each time segment is calculated based on its expected maintenance duration and the availability of technical personnel. For example, the allocation probability of Task 1 in the 10:00-10:15 AM time segment is 0.08, while the allocation probability in the 12:00-12:15 PM time segment is only 0.02, reflecting the difference in resource utilization efficiency at different times.

[0059] The scheduling optimization algorithm is implemented using an improved ant colony algorithm. Path search for the execution order of maintenance tasks is performed based on the transition probability matrix in the outer layer encoding. Taking the five maintenance tasks mentioned above as an example, assuming the optimal path obtained after 100 iterations is 4-1-3-5-2, it means the maintenance work should be executed in the order of task 4, task 1, task 3, task 5, and task 2. Simultaneously, the execution time of each maintenance task is determined based on the time allocation probability matrix in the inner layer encoding. For example, for task 4, the algorithm might determine its optimal execution time as 8:30-9:45 AM; for task 1, the optimal execution time is 10:00-11:15 AM, and so on, completing the time allocation for all tasks.

[0060] In actual scheduling, various constraints need to be considered. The time window constraint and resource constraint in the predicted maintenance indicators are used as temperature parameters to locally adjust the maintenance task allocation scheme obtained from the two-layer coding structure search. The time window constraint means that a specific maintenance task must be completed within a specified time frame; for example, task 3 must be completed before 2:00 PM on the same day. Resource constraints refer to limiting factors such as the number of available maintenance personnel and spare parts. The introduction of the temperature parameter allows the algorithm to appropriately accept suboptimal solutions while satisfying the constraints, avoiding getting trapped in local optima. When the allocation scheme of a maintenance task violates the constraints, it is adjusted to the nearest feasible time point, and the objective function value is recalculated. The objective function comprehensively considers maintenance timeliness, resource utilization, and constraint satisfaction, and can be expressed as a weighted sum of the three. The weights can be set according to actual needs, for example, maintenance timeliness weight 0.5, resource utilization weight 0.3, and constraint satisfaction weight 0.2. Based on the adjusted objective function value, the initial pheromone distribution is updated to increase the pheromone concentration on high-quality paths and decrease the pheromone concentration on low-quality paths.

[0061] To obtain the globally optimal maintenance task allocation scheme, a scheduling optimization algorithm based on a two-layer coding structure is repeatedly executed until a preset number of iterations, such as 200, is reached. During the iteration process, the algorithm continuously optimizes the execution order and timing of maintenance tasks, gradually improving the quality of the scheduling scheme. The final output maintenance task allocation scheme includes the execution order and specific timing of each instrument maintenance task, for example, "Task 4 will be executed from 8:30 to 9:45, Task 1 from 10:00 to 11:15, Task 3 from 13:00 to 14:30, Task 5 from 15:00 to 16:15, and Task 2 from 16:30 to 17:45."

[0062] This embodiment calculates and predicts maintenance indicators based on real-time data collected by sensors, and combines this with a two-layer coding structure for intelligent scheduling. This significantly improves the efficiency and accuracy of instrument maintenance in the Internet of Things environment. It not only enables real-time monitoring and prediction of instrument status, but also automatically generates the optimal maintenance scheduling plan based on the monitoring results. This greatly improves the availability and lifespan of laboratory instruments, reduces maintenance costs, and provides effective technical support for the full life cycle management of laboratory instruments.

[0063] like Figure 2 The flowchart shown is a preview of the predictive maintenance task scheduling optimization algorithm.

[0064] In one optional implementation, the path search for the execution order of maintenance tasks is performed based on the transition probability matrix, and the execution time of the maintenance tasks is determined based on the time allocation probability matrix, further including: In the outer path encoding, the execution order position of the next maintenance task is selected based on the transition probability matrix; In the inner time encoding, the execution time of the maintenance task corresponding to the execution order position is determined based on the time allocation probability matrix; The execution result of the maintenance task at the specified execution time is fed back to the outer path encoding to update the transition probability between the maintenance task and other maintenance tasks in the transition probability matrix.

[0065] In one specific implementation, the outer path encoding selects the execution order of the next maintenance task based on a transition probability matrix. The transition probability matrix is ​​a two-dimensional array representing the transition relationships between different maintenance tasks, where each element represents the probability of transitioning from one maintenance task to another. Taking five main analytical instruments in a laboratory as an example, these instruments achieve real-time acquisition and transmission of status data via an IoT module. The transition probability matrix has a dimension of 5×5, with each row representing the probability distribution of selecting the next maintenance instrument starting from the current instrument. Assuming the maintenance of the first analytical instrument has been completed, the transition probability matrix constructed based on the status monitoring data shows: the probability of transitioning to the second analytical instrument is 0.42, the probability of transitioning to the third testing instrument is 0.28, the probability of transitioning to the fourth measuring instrument is 0.18, and the probability of transitioning to the fifth monitoring device is 0.12. The calculation of the transition probabilities comprehensively considers factors such as instrument health status, usage frequency, maintenance urgency, and importance weight. The next maintenance target is selected from these probabilities using a roulette wheel selection algorithm. Specifically, a random number between 0 and 1 (0.37) is generated. Since this value is less than the transition probability of 0.42 from the first analyzer to the second, the second analyzer is selected as the next instrument to be maintained. This probability-based selection mechanism ensures that instruments with high maintenance urgency receive a higher selection probability, while introducing randomness to avoid getting trapped in local optima.

[0066] After determining the maintenance sequence, the inner time coding uses a time allocation probability matrix to determine the execution time of the maintenance task corresponding to the execution sequence position. The time allocation probability matrix represents the probability distribution of executing the maintenance task at different time points. Taking an 8-hour workday as an example, the work time is divided into 16 30-minute time segments. For the selected second analytical instrument, a 1×16 time allocation probability matrix is ​​established. This matrix is ​​calculated based on technician availability, equipment usage plans, and maintenance resource status. For example, the allocation probability for 9:00-9:30 AM is 0.12, for 9:30-10:00 AM it is 0.15, for 10:00-10:30 AM it is 0.18, while for 12:00-12:30 PM it is only 0.03. Using the same roulette wheel selection method, a random number of 0.45 is generated, determining the maintenance start time for the second analytical instrument to be 10:00 AM. Considering that the maintenance of this instrument is expected to take 90 minutes, combined with the time segment division, its maintenance period is 10:00-11:30 AM. The time allocation takes into account the expected duration of the maintenance task, the workload balancing of technical personnel, and the instrument usage plan, ensuring that the maintenance task is completed efficiently without affecting normal experiments.

[0067] After completing the maintenance task for the second analytical instrument, the execution results are fed back to the outer path encoding, updating the transition probability matrix between this instrument and other instruments. The feedback information includes multiple dimensions: anomalies discovered during maintenance, the degree of performance improvement after maintenance, the deviation between actual and planned maintenance time, and maintenance resource consumption. IoT sensors continuously collect and compare instrument status data before and after maintenance. For example, after maintenance, the stability of the second analytical instrument improved by 18%, background noise decreased by 25%, and accuracy increased by 12%. Simultaneously, the IoT platform detected abnormal fluctuations in key parameters of the third detection instrument, indicating a deterioration in its condition. Based on this feedback information, the transition probability matrix is ​​updated to reflect the transition probabilities from the second analytical instrument to other instruments: the transition probability from the second analytical instrument to the third detection instrument increases from 0.28 to 0.45; the transition probability to the fourth measuring instrument remains unchanged at 0.18; the transition probability to the fifth monitoring device decreases from 0.12 to 0.07; and the transition probability to the first analytical instrument decreases from 0.42 to 0.30.

[0068] The calculation process for updating the transfer probability includes four main steps: state assessment, weight calculation, transfer tendency calculation, and probability normalization. State assessment calculates the health index of each instrument based on IoT data, taking into account deviations in parameters such as temperature, pressure, current, and flow rate. For example, if the health index of the third testing instrument drops from 85 to 67, it indicates a significant deterioration in its condition. Weight calculation considers the importance and usage frequency of the instruments, assigning basic weights to different instruments. For instance, the importance weight of the third testing instrument is 0.25, higher than other instruments. Transfer tendency calculation combines the health index with the importance weight to derive the transfer tendency value from the current instrument to each target instrument. Finally, probability normalization ensures that the sum of all transfer probabilities is 1, forming an effective probability distribution.

[0069] The updated transition probability matrix is ​​used to select the next maintenance task, continuing the task selection process in the outer path coding. Based on the new transition probabilities, the third testing instrument is selected as the next maintenance target. Then, the time allocation step of the inner time coding is executed again to determine the maintenance time for the third testing instrument as 13:30-15:30. This cycle repeats until all instrument tasks requiring maintenance are completed or the preset scheduling cycle end condition is met.

[0070] The combination of a dual-layer coding structure and a feedback mechanism enables maintenance scheduling to dynamically adapt to changes in instrument status, achieving intelligent and adaptive maintenance task scheduling. The outer path coding layer addresses the questions of "what to maintain" and "maintenance sequence," while the inner time coding layer addresses the question of "when to maintain." The feedback mechanism allows the entire system to continuously learn and optimize, forming a closed-loop control. The application of IoT technology automates and enables real-time acquisition and transmission of instrument status data, providing accurate data support for the dual-layer coding structure and making maintenance decisions more precise and timely.

[0071] The instrument lifecycle management system based on the Internet of Things (IoT) according to embodiments of the present invention includes: The first unit is used to bind RFID tags to instruments to obtain unique identification information, collect instrument operating parameters, and establish a multi-dimensional data structure. The second unit is used to perform time-series clustering analysis on the multidimensional data structure, identify abnormal operating patterns of the instrument, construct an equipment group association diagram, calculate fault propagation characteristics based on the fault state transfer between nodes in the equipment group association diagram, and form an equipment operating status assessment result. The third unit is used to construct the instrument's performance degradation curve based on the equipment operation status assessment results by extracting multi-head self-attention features and optimizing fitting functions, and to perform probability calculations by combining the fault propagation features to generate predictive maintenance indicators. The fourth unit is used to generate a maintenance task allocation scheme based on the predicted maintenance indicators and using a scheduling optimization algorithm with the goals of maintenance timeliness and resource utilization. The fifth unit is used to perform maintenance work based on the maintenance task allocation results, store the maintenance process and result data in the maintenance knowledge base, and update the predicted maintenance indicators and the maintenance task allocation results through online analysis.

[0072] A third aspect of the present invention provides an electronic device, comprising: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0073] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0074] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for managing the entire lifecycle of instruments based on the Internet of Things, characterized in that, include: Bind RFID tags to instruments to obtain unique identification information, collect instrument operating parameters, and establish a multi-dimensional data structure; Temporal clustering analysis is performed on the multidimensional data structure to identify abnormal operating patterns of the instrument, and an equipment group association graph is constructed. Based on the fault state transition between nodes in the equipment group association graph, fault propagation characteristics are calculated to form equipment operating status assessment results. Based on the equipment operation status assessment results, the performance degradation curve of the instrument is constructed by extracting multi-head self-attention features and optimizing fitting functions. Probability calculation is performed in combination with the fault propagation features to generate predictive maintenance indicators. Based on the predicted maintenance indicators, a scheduling optimization algorithm is used to generate a maintenance task allocation scheme with the goals of maintenance timeliness and resource utilization. Maintenance work is performed based on the maintenance task allocation results, and the maintenance process and result data are stored in the maintenance knowledge base. The predicted maintenance indicators and the maintenance task allocation results are updated through online analysis.

2. The method according to claim 1, characterized in that, Performing time-series clustering analysis on the multidimensional data structure to identify abnormal instrument operation patterns and constructing an equipment group association graph includes: Temporal feature extraction is performed on multidimensional data structures, extracting statistical features, frequency domain features, and time-frequency features from the multidimensional data structures. The optimal feature subset is determined based on the feature importance weights, and a dimension-reduced feature matrix is ​​generated. Based on the dimensionality-reduced feature matrix, dynamic time warping is used to calculate the similarity of time series sequences, a dynamic time warping distance matrix is ​​constructed, and the time series clustering results are obtained by iteratively calculating the similarity between samples and cluster centers; The time-series clustering results are compared with a preset normal operation mode benchmark to calculate the clustering deviation value. Based on the comparison result of the clustering deviation value and the preset anomaly judgment threshold, an abnormal mode of instrument operation is generated. Based on the abnormal operation patterns of the instruments, a device group association graph is constructed. The device operation status parameters are extracted to calculate the correlation coefficient between devices, the association strength threshold is determined to establish node connection relationships, and the edge weights are calculated based on the device physical location, operation parameters and the abnormal patterns to generate a weighted device group association graph.

3. The method according to claim 1, characterized in that, Based on the fault state transition between nodes within the equipment group association diagram, the fault propagation characteristics are calculated to form the equipment operating status assessment results, including: Within a preset time window, the number of fault state transitions between adjacent nodes in the device group association diagram is counted. The number of fault state transitions is divided by the length of the time window to obtain the fault transition probability between nodes, forming a fault propagation probability matrix. By gradually updating the propagation distance between nodes, the shortest propagation path and propagation time delay between each pair of nodes in the fault propagation probability matrix are calculated. Nodes with a fault transfer probability greater than a preset probability threshold are marked as key propagation nodes, and a fault propagation path map is generated. The number of connecting edges of each node in the fault propagation path graph is counted. The number of connecting edges is multiplied by the fault propagation probability of each edge to calculate the node influence degree. The ratio of the fault propagation distance between nodes to the sampling interval is calculated to obtain the fault spread rate. Areas where the node influence degree and fault spread rate exceed the corresponding thresholds are marked as propagation danger zones. The distance from the edge nodes of the propagation danger zone to the key propagation nodes is calculated to determine the fault impact range. The mean and variance of equipment operating parameters are calculated to obtain the single equipment status index, the correlation coefficient between parameters of equipment is calculated to obtain the group coordination index, the risk index is determined by weighting the node influence degree, diffusion speed and fault impact range, and the weighted average of the single equipment status index, group coordination index and risk index is used to obtain the equipment operating status assessment result.

4. The method according to claim 1, characterized in that, Based on the equipment operation status assessment results, the performance degradation curve of the instrument is constructed by multi-head self-attention feature extraction and fitting function optimization, including: The device operation status evaluation results are input into the multi-head self-attention layer, which includes a query-key-value calculation unit. The query-key-value calculation unit captures features at different time scales and determines the attention features. The attention features are input into the temporal position coding layer, and the relative position of the attention features is represented by sine and cosine functions to determine the temporal dependency and generate position coding features. The location encoding features are transformed by dimensional mapping to generate intermediate features. The intermediate features are then passed through an activation function to obtain transformed features. The transformed features are then superimposed with the location encoding features to establish short-circuit connection paths that preserve the original feature information, resulting in fused features. Based on the fusion features, a state change rate sequence is constructed, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features. Based on the time interval and amplitude difference of the state change points, the attenuation feature groups are grouped to construct attenuation feature groups, which are then stored in the attenuation feature database, and the optimal fitting function is selected. Calculate the mean squared error between the multi-scale performance characteristics and the optimal fitting function, construct a gradient descent matrix, adjust the fitting coefficients of the optimal fitting function according to the gradient descent matrix, repeat the iteration until the mean squared error is less than a preset termination threshold, and obtain the instrument performance degradation curve.

5. The method according to claim 4, characterized in that, Based on the fused features, a state change rate sequence is constructed, state change points are determined, and periodic change features and long-term evolution features of the state change points are extracted to generate multi-scale performance features, including: The difference values ​​between adjacent time points are calculated for the fused features according to the time series to construct a state change rate sequence. The difference values ​​are then calculated for the state change rate sequence to construct a performance change acceleration sequence. The state change point is determined based on the local maxima and local minima of the state change rate sequence and the zero point of the performance change acceleration sequence. The fluctuation patterns of the state change points within different time windows are extracted to construct a state fluctuation feature sequence; periodic change features are extracted based on the recurrence frequency and amplitude changes of the state fluctuation feature sequence; and a moving average is applied to the state change points to extract the baseline trend of state changes as a long-term evolution feature. By weighting and combining periodic change characteristics and long-term evolution characteristics, multi-scale performance characteristics are generated.

6. The method according to claim 1, characterized in that, The scheduling optimization algorithm includes: A two-layer coding structure for maintenance tasks is constructed based on predictive maintenance indicators. The two-layer coding structure includes an outer path coding for determining the execution order of maintenance tasks and an inner time coding for determining the execution time of maintenance tasks. In the outer path encoding, the initial pheromone distribution between maintenance tasks is set, and the transition probability matrix between maintenance tasks is constructed based on the maintenance timeliness index; A time allocation space for maintenance tasks is established in the inner time encoding, and a time allocation probability matrix for maintenance tasks is constructed based on the resource utilization index. The path search for the execution order of maintenance tasks is performed based on the transition probability matrix, and the execution time of maintenance tasks is determined based on the time allocation probability matrix. Using the time window constraint and resource constraint in the predicted maintenance index as temperature parameters, the maintenance task allocation scheme obtained by the search of the dual-layer coding structure is locally adjusted, and the initial pheromone distribution is updated based on the adjusted objective function value. Based on the two-layer coding structure, the scheduling optimization algorithm is repeatedly executed until the preset number of iterations is reached, and a maintenance task allocation scheme is output.

7. The method according to claim 6, characterized in that, The path search for the execution order of maintenance tasks is based on the transition probability matrix, and the execution time of maintenance tasks is determined based on the time allocation probability matrix. This also includes: In the outer path encoding, the execution order position of the next maintenance task is selected based on the transition probability matrix; In the inner time encoding, the execution time of the maintenance task corresponding to the execution order position is determined based on the time allocation probability matrix; The execution result of the maintenance task at the specified execution time is fed back to the outer path encoding to update the transition probability between the maintenance task and other maintenance tasks in the transition probability matrix.

8. An Internet of Things-based instrument lifecycle management system, used to implement the method of any one of claims 1-7, characterized in that, include: The first unit is used to bind RFID tags to instruments to obtain unique identification information, collect instrument operating parameters, and establish a multi-dimensional data structure. The second unit is used to perform time-series clustering analysis on the multidimensional data structure, identify abnormal operating patterns of the instrument, construct an equipment group association diagram, calculate fault propagation characteristics based on the fault state transfer between nodes in the equipment group association diagram, and form an equipment operating status assessment result. The third unit is used to construct the instrument's performance degradation curve based on the equipment operation status assessment results by extracting multi-head self-attention features and optimizing fitting functions, and to perform probability calculations by combining the fault propagation features to generate predictive maintenance indicators. The fourth unit is used to generate a maintenance task allocation scheme based on the predicted maintenance indicators and using a scheduling optimization algorithm with the goals of maintenance timeliness and resource utilization. The fifth unit is used to perform maintenance work based on the maintenance task allocation results, store the maintenance process and result data in the maintenance knowledge base, and update the predicted maintenance indicators and the maintenance task allocation results through online analysis.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images