Intelligent laboratory full-process collaborative management system and method based on multi-dimensional data fusion

Abstract

The invention provides an intelligent laboratory full-process collaborative management system and method based on multi-dimensional data fusion, and the method comprises the steps: obtaining an original data stream from laboratory instrument equipment, carrying out the format recognition and conversion of heterogeneous instrument data through a standard protocol adapter, and generating a first data set in a unified format; aiming at the first data set, executing automatic analysis by adopting data middleware, and integrating into a structured second data set based on a metadata rule of field mapping and unit conversion; according to the second data set, a multi-source data fusion model is constructed, data are classified according to instrument types through batch processing and aggregated according to timestamps, and a fused third data set is generated; aiming at the initial process configuration, adjusting a node sequence and a parameter threshold value by using a visual design tool, and generating an optimized process configuration adaptive to the diversified scene; according to the optimized process configuration, process scheduling is executed through a dynamic process engine, execution time and resource occupation are monitored in real time, and a process execution log is generated.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a smart laboratory full-process collaborative management system and method based on multidimensional data fusion. Background Art

[0002] Problem background:

[0003] Laboratory management is a crucial pillar of modern scientific research and testing, directly impacting data quality, efficiency, and compliance. Its core lies in achieving resource optimization, process standardization, and quality control. However, existing laboratory management methods have significant limitations. Distributed data storage leads to information silos, manual operations rely on paper records, resulting in inefficient and difficult traceability processes, lagging quality control and the lack of real-time early warning, and a lack of dynamic monitoring of equipment and consumables. These shortcomings collectively hinder the digital transformation and efficient operation of laboratories.

[0004] Specifically, current laboratory management faces the core challenges of multi-source data integration and dynamic process optimization. First, the wide variety of instruments and equipment, heterogeneous data formats, and the lack of unified middleware for automatic parsing and integration lead to inefficient data collection. Second, testing processes rely on fixed templates, making them difficult to adapt to diverse scenarios and lacking flexible process design tools to support dynamic adjustments. These unresolved technical factors have led to unique challenges such as data silos, process rigidity, and quality control delays, severely impacting laboratory operational efficiency and compliance.

[0005] Therefore, how to design a middleware that supports automatic parsing and integration of heterogeneous instrument data and develop a visual design tool based on a dynamic process engine to achieve seamless integration of multi-source data and flexible configuration of detection processes has become a key issue in the digital transformation of laboratory management. Summary of the Invention

[0006] The present invention provides a smart laboratory full-process collaborative management system and method based on multidimensional data fusion, which mainly includes:

[0007] Obtaining raw data streams from laboratory instruments and equipment, performing format recognition and conversion on heterogeneous instrument data using a standard protocol adapter, and generating a first data set in a unified format;

[0008] For the first data set, data middleware is used to perform automatic parsing and integrate it into a structured second data set based on metadata rules of field mapping and unit conversion;

[0009] Based on the second data set, a multi-source data fusion model is constructed, and the data is classified by instrument type and aggregated by timestamp through batch processing to generate a fused third data set;

[0010] For the third data set, check the data integrity. If the missing rate is lower than the preset threshold and the format is consistent, load the detection process template through the dynamic process engine to generate the initial process configuration. If it is not satisfied, return to the second data set to supplement the missing data and re-parse;

[0011] For the initial process configuration, use visual design tools to adjust the node order and parameter thresholds to generate optimized process configurations that adapt to diverse scenarios;

[0012] Based on the optimized process configuration, the dynamic process engine is used to execute process scheduling, monitor execution time and resource usage in real time, and generate process execution logs;

[0013] Based on the process execution log, a statistical analysis algorithm is used to detect quality control indicators. When the indicators deviate from the preset threshold, an early warning signal is generated and the process configuration parameters are updated and optimized. If the threshold is met, a test result report is generated;

[0014] Based on the detection result report, the metadata mapping rules of the multi-source data fusion model are updated, the field priority is adjusted and applied to the second data set, and the data integration efficiency is continuously improved.

[0015] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a flow chart of a smart laboratory full-process collaborative management system and method based on multidimensional data fusion of the present invention.

[0017] Figure 2 This is a schematic diagram of a smart laboratory full-process collaborative management system and method based on multi-dimensional data fusion according to the present invention.

[0018] Figure 3 This is another schematic diagram of a smart laboratory full-process collaborative management system and method based on multi-dimensional data fusion according to the present invention. DETAILED DESCRIPTION

[0019] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] like Figure 1-3 In this embodiment, a smart laboratory full-process collaborative management system and method based on multi-dimensional data fusion may specifically include:

[0021] Step S101 : obtaining an original data stream from laboratory instruments and equipment, performing format recognition and conversion on heterogeneous instrument data using a standard protocol adapter, and generating a first data set in a unified format.

[0022] Developed heterogeneous instrument data acquisition middleware to support automatic parsing of data from mainstream devices. The middleware includes built-in standard protocol adapters to perform format recognition and conversion for different instrument types and data formats. The middleware periodically acquires data streams based on the data acquisition frequency and parses them according to data parsing rules. Parsed data is verified using a data validation mechanism to ensure data integrity. Validated data is converted into a unified format using data conversion logic. The converted data is stored in a designated location and synchronized according to the data synchronization frequency. The middleware implements data security policies to protect data privacy and controls data access by setting data access permissions.

[0023] For example, in a laboratory environment, when obtaining raw data streams through instruments and equipment, a standard protocol adapter is first used to perform format recognition and conversion on heterogeneous instrument data.

[0024] For example, when acquiring data from a mass spectrometer and a chromatograph, the mass spectrometer outputs data in XML format, while the chromatograph outputs data in CSV format. Using an adapter, the mass spectrometer's XML data is parsed into structured data containing mass-to-charge ratio (m / z) and intensity values, while the chromatograph's CSV data is parsed into structured data containing retention time and peak area. The adapter uses regular expressions and an XML parser to identify the data, ensuring its accuracy and completeness. Next, the parsed data is converted using a unified data model.

[0025] For example, data from a mass spectrometer is converted into a JSON format containing m / z values, intensity values, and timestamps, while data from a chromatograph is converted into a JSON format containing retention times, peak areas, and timestamps. During the conversion process, a hashing algorithm is used to verify the data to ensure it has not been tampered with. Ultimately, the resulting first data set contains data from all instruments in a unified format, facilitating subsequent data analysis and processing.

[0026] For example, a mass spectrometer data set contains records with an m / z value of 100.5, an intensity value of 5000, and a timestamp of 2023-10-01T12:00:00. A chromatograph data set contains records with a retention time of 5.2 minutes, a peak area of ​​12000, and a timestamp of 2023-10-01T12:00:00. These data sets are provided to the data analysis module through a unified interface for further processing and analysis.

[0027] In step S102 , the first data set is automatically parsed using data middleware, and integrated into a structured second data set based on metadata rules of field mapping and unit conversion.

[0028] Data middleware performs automatic parsing, integrating the first data set into a structured second data set based on metadata rules for field mapping and unit conversion. If the data format is text, the field name and field type are parsed; if the data format is binary, the parsing rules are determined based on the device type. Data units are unified based on the unit type and conversion rules. Data source identifiers and timestamps are used to associate device status with experimental records. The parsed data is written to a structured target, completing the data integration.

[0029] Exemplarily, during the data processing process, the first data set is first automatically parsed by the data middleware to identify key fields therein.

[0030] For example, suppose the first dataset contains a "temperature" field whose original data is "25°C." The middleware uses a regular expression to extract the numeric value "25" and the unit "°C." Next, based on predefined field mapping rules, the "temperature" field is mapped to the "ambient temperature" field in the second dataset. Simultaneously, according to the unit conversion rules, "°C" is converted to "K" using the formula "K = °C + 273.15." Therefore, "25°C" is converted to "298.15K." Furthermore, suppose the first dataset also contains a "humidity" field whose original data is "60%." The middleware uses the percentage conversion rule to convert it to a decimal form of "0.6" and maps it to the "relative humidity" field in the second dataset. Through this series of operations, the first dataset is integrated into a structured second dataset, where the "ambient temperature" is "298.15K" and the "relative humidity" is "0.6," ensuring data accuracy and consistency.

[0031] Step S103 : constructing a multi-source data fusion model based on the second data set, classifying the data by instrument type and aggregating them by timestamp through batch processing, and generating a fused third data set.

[0032] Build a time series database cluster, divide data into shards by instrument type, design a shard key that includes the device fingerprint field, establish an aggregation strategy for minute-level time windows, and configure stream processing jobs to clean and transform the raw data stream.

[0033] For example, when building a multi-source data fusion model, it is first necessary to extract key fields from the second data set. For example, the instrument type field includes "temperature sensor", "hygrometer", "barometer", etc., and the timestamp field is accurate to the millisecond level, such as "2023-05-12 14:30:45.678". The MapReduce framework is used for batch processing, in which the Mapper stage performs hash partitioning based on the instrument type field and assigns data of the same type to the same Reducer. For example, all "temperature sensor" data will be assigned to Reducer1, and its data format is [device ID, timestamp, temperature value].

[0034] ["T001","2023-05-1214:30:00.000",25.3]. The Reducer stage uses a time window aggregation algorithm, sets a 5-minute time window, and calculates the mean of the data in each window. For example, in the 14:30-14:35 time window, the three temperature readings [25.3,25.7,25.5] are calculated to get an aggregate value of 25.5. The 3σ principle is used for outlier detection. When a reading exceeds the mean ±3 times the standard deviation range, it is automatically eliminated. The fused third data set is stored in JSON format. Each record contains fields such as instrument type, aggregation timestamp, and aggregation value. For example, {"instrument_type":"temperature sensor","timestamp":"2023-05-1214:30:00",

[0035] "value":25.5, "unit":"℃"}. The data quality control module records the amount of original data, valid data, and abnormal data in each time window, such as "Original data amount: 15, Valid data amount: 14, Abnormal data amount: 1". This metadata is stored along with the fusion results.

[0036] Step S104: Check the data integrity of the third data set. When the missing rate is lower than the preset threshold and the format is consistent, load the detection process template through the dynamic process engine to generate the initial process configuration. If it is not satisfied, return to the second data set to supplement the missing data and re-parse.

[0037] Check the data integrity of the third data set to determine whether the missing rate is lower than the preset threshold and the format is consistent. If the conditions are met, load the detection process template through the dynamic process engine to generate the initial process configuration. If not, return to the second data set to supplement the missing data and re-parse. Use the built-in 19-category quality control model to perform real-time quality control on the data, record all data modification trajectories, and support traceability in three dimensions: personnel, operations, and impact results. Use RFID technology to track the entire life cycle of hazardous chemicals, and link inventory thresholds to the procurement system. Temperature and humidity sensor data are written into the detection record in real time, and the experiment is automatically suspended if the standard is exceeded. Establish a multi-dimensional capability matrix to automatically match detection tasks with personnel qualifications. Direct instrument connection is achieved through the HTTP protocol text protocol, automatically recording the frequency of use and triggering calibration reminders. Built-in standard detection method library, automatic synchronization of experimental templates with version changes. Support dynamic adjustment of the detection process, realize electronic signature nesting from entrustment registration to report issuance, and embed blockchain timestamps in key nodes.

[0038] Exemplarily, in the third data set, the missing rate is first calculated by the data integrity check algorithm. Assuming that the preset threshold is 5%, if the missing rate is 3%, which is lower than the threshold, the data format consistency is checked again, and regular expressions are used to match all data items to ensure uniform format. If the format is consistent, the dynamic process engine will load the detection process template and generate the initial process configuration. For example, the configuration includes steps such as data cleaning, feature extraction, and model training. If the missing rate is higher than 5% or the format is inconsistent, the system will return to the second data set and use a data completion algorithm such as the KNN interpolation method to supplement the missing data. Assuming that the K value is 3, the nearest neighbor mean of each missing value is calculated to fill it in, and then the data set is re-parsed to ensure data integrity and format consistency before generating the process configuration again. The entire process is implemented through automated scripts to ensure the efficiency and accuracy of data processing.

[0039] Step S105 : For the initial process configuration, use a visual design tool to adjust the node sequence and parameter thresholds to generate an optimized process configuration that adapts to diverse scenarios.

[0040] The process node topology diagram is displayed in the visual designer, and the node type and connection relationship are marked. If the node type is instrument operation, the device type and data collection frequency are associated. If the node type is quality control inspection, the quality control rules and exception handling logic are embedded. According to the detection scenario classification, the node sequence and parameter threshold range are adjusted, the personnel qualification requirements and material properties are matched, the environmental condition monitoring points are set, the task priority and report generation rules are configured, and the optimized process configuration plan is generated.

[0041] For example, during the initial process configuration, the node order is first adjusted using a visual design tool. For example, the data preprocessing node is placed before the feature extraction node to ensure that the data has been cleaned and standardized before entering the feature extraction stage. Next, to optimize parameter thresholds, a grid search algorithm is used to tune key model parameters. For example, the C parameter of the support vector machine (SVM) is adjusted from the default value of 1.0 to 10.0 to improve the model's classification performance. During the optimization process, the model is evaluated using cross-validation to ensure that the adjusted parameters maintain stable performance across different datasets. In addition, to meet the needs of diverse scenarios, an adaptive threshold adjustment mechanism is introduced. By monitoring the model output in real time, the classification threshold is dynamically adjusted. For example, in a binary classification problem, the threshold is adjusted from 0.5 to 0.6 to reduce the false positive rate. Finally, multiple optimized models are combined using ensemble learning techniques. For example, the prediction results of the random forest and gradient boosting trees are combined using a weighted average method, with weights set to 0.4 and 0.6, respectively, to further improve overall prediction accuracy. The entire optimization process configuration is implemented through automated scripts to ensure that appropriate process configurations can be quickly generated in different scenarios.

[0042] Step S106 , according to the optimized process configuration, process scheduling is performed through the dynamic process engine, execution time and resource usage are monitored in real time, and a process execution log is generated.

[0043] Establish a multi-dimensional capability matrix, automatically match detection tasks with personnel qualifications, realize direct connection of instruments through HTTP protocol text protocol, automatically record usage frequency and trigger calibration reminders, use RFID technology to realize full life cycle tracking of hazardous chemicals, inventory threshold linkage procurement system, built-in standard detection method library, automatic synchronization of experiment templates with version changes, real-time writing of temperature and humidity sensor data into detection records, automatic suspension of experiments when exceeding standards, intelligent task routing engine supports dynamic adjustment of detection process, paperless whole process from entrustment registration to report issuance, realize electronic signature nesting, key nodes are embedded in blockchain timestamps, establish 19 types of quality control models, audit tracking matrix records all data modification trajectories, and supports three-dimensional traceability based on personnel-operation-impact results.

[0044] For example, in optimizing process configuration, tasks are first scheduled through the dynamic process engine. For example, a priority-based scheduling algorithm is used to sort tasks from high to low priority, ensuring that high-priority tasks are executed first. For example, if there are three tasks in the system with priorities of 9, 5, and 3, the scheduling algorithm will prioritize the task with priority 9. When monitoring execution time and resource usage in real time, the system records the start and end times of each task and calculates its execution duration.

[0045] For example, Task A starts at 10:00 and ends at 10:05, with an execution time of 5 minutes. At the same time, the system monitors the CPU and memory usage. Assume that the CPU usage is 80% and the memory usage is 512MB during the execution of Task A. By analyzing this data, the system can determine whether there is a resource bottleneck. For example, if a task takes too long to execute and the CPU usage is continuously above 90%, the resource allocation of the task may need to be optimized. Finally, the system generates a process execution log, which records the execution status, time, resource usage, and other information of each task. The log format can be JSON, containing fields such as task ID, start time, end time, CPU usage, and memory usage.

[0046] For example, the log of task A is {“task_id”:“A”, “start_time”:

[0047] "10:00", "end_time": "10:05", "cpu_usage": 80, "memory_usage": 512}. Using these logs, the system can further analyze process execution efficiency, optimize resource allocation, and improve overall performance.

[0048] Step S107: Statistical analysis algorithms are used to detect quality control indicators in the process execution log. When the indicators deviate from the preset threshold, an early warning signal is generated and the optimization process configuration parameters are updated. If the threshold is met, a detection result report is generated.

[0049] A process log analysis engine was established, and a sliding time window algorithm was used to process time series data to calculate task turnover rate and equipment idle rate indicators.

[0050] For example, in the process execution log analysis, the key indicators in the log (such as average processing time and error rate) are first counted in real time through the sliding window algorithm. The window size is 100 log records and the step size is 10. The mean and standard deviation of the indicators in each window are calculated.

[0051] For example, when the average processing time within a window exceeds a preset threshold of 500 milliseconds (standard deviation ±50 milliseconds), an isolation forest algorithm is used to detect outliers, with an anomaly score threshold set to 0.65. If anomalies are detected in three consecutive windows, an alert is triggered. After the alert is generated, the system automatically calls a Bayesian optimization algorithm to adjust process parameters, such as increasing the number of thread pool cores from 8 to 12. Gradient descent (learning rate 0.01, 100 iterations) is then used to verify whether the metrics return to normal under the new parameters. For processes that meet the threshold, the system generates a report containing a KPI trend chart (e.g., a 15% decrease in error rate over 30 days) and statistical test results (p-value < 0.05). The reported data is then reduced in dimension using principal component analysis and stored in a database, retaining the top three principal components (cumulative variance contribution ≥ 85%). All operations are implemented through an event-driven architecture, with structured data (JSON format, containing fields such as timestamps, metric values, and confidence intervals) transmitted between modules via message queues, ensuring real-time and traceable analysis.

[0052] Step S108: Based on the detection result report, the metadata mapping rules of the multi-source data fusion model are updated, the field priority is adjusted and applied to the second data set, so as to continuously improve the data integration efficiency.

[0053] Multi-source data fusion technology is used to match the collected test result data with the metadata mapping rule library. If the test result data is inconsistent with the existing rules, the rule update mechanism is triggered, the field priority is adjusted, the metadata mapping rule library is updated, and the updated rules are applied to the second data set. Through the intelligent task routing engine, the updated data set is associated with the test process. Combined with the personnel qualification matrix, the test tasks are automatically assigned, and the task assignment results are compared with the quality control model. If the quality control model detects abnormal data, a real-time quality control dashboard warning is triggered, and the warning information is associated with the audit tracking matrix to record the data modification trajectory. Through the dynamic process engine, the updated data set is synchronized with the test process template to continuously improve data integration efficiency.

[0054] For example, based on the detection result report showing that data source A's field coverage was 85% and data source B's was 72%, a weighted fusion algorithm was used to adjust the metadata mapping rules, setting the field weights for data source A to 0.6 and for data source B to 0.4. For conflicting fields such as "product number," a priority coverage strategy was implemented based on the confidence score (0.92 for data source A and 0.87 for data source B). When applied to the second dataset, a modified Jaccard similarity algorithm was used to calculate field matching, with a fusion threshold of 0.75 set. Automatic merging was triggered when the similarity for the "customer address" field reached 0.82. A TF-IDF-based text feature extraction method was used to process unstructured data, with a feature dimension set to 256. A random forest classifier (with 100 decision trees) was used to intelligently map the "product category" field, achieving an accuracy rate of 89.3%. Real-time quality monitoring was implemented during the data integration process. When the missing value ratio exceeded 15%, a KNN (k=5)-based missing value imputation algorithm was automatically triggered to ensure data integrity remained above 93%. For time series data, a dynamic time warping algorithm (with a 7-day window) is used to align timestamps from different sources, keeping the error within ±2 hours. By continuously optimizing these parameters and algorithms, the system's ability to process data records per hour has increased from 120,000 to 180,000, reducing processing latency by 22%.

[0055] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A smart laboratory full-process collaborative management system and method based on multi-dimensional data fusion, characterized by: The method comprises: Obtaining raw data streams from laboratory instruments and equipment, performing format recognition and conversion on heterogeneous instrument data using a standard protocol adapter, and generating a first data set in a unified format; For the first data set, data middleware is used to perform automatic parsing and integrate it into a structured second data set based on metadata rules of field mapping and unit conversion; Based on the second data set, a multi-source data fusion model is constructed, and the data is classified by instrument type and aggregated by timestamp through batch processing to generate a fused third data set; For the third data set, check the data integrity. If the missing rate is lower than the preset threshold and the format is consistent, load the detection process template through the dynamic process engine to generate the initial process configuration. If it is not satisfied, return to the second data set to supplement the missing data and re-parse; For the initial process configuration, use visual design tools to adjust the node order and parameter thresholds to generate optimized process configurations that adapt to diverse scenarios; Based on the optimized process configuration, the dynamic process engine is used to execute process scheduling, monitor execution time and resource usage in real time, and generate process execution logs; Based on the process execution log, a statistical analysis algorithm is used to detect quality control indicators. When the indicators deviate from the preset threshold, an early warning signal is generated and the process configuration parameters are updated and optimized. If the threshold is met, a test result report is generated; Based on the detection result report, the metadata mapping rules of the multi-source data fusion model are updated, the field priority is adjusted and applied to the second data set, and the data integration efficiency is continuously improved.

2. The method according to claim 1, characterized in that The method of obtaining a raw data stream from laboratory instruments and equipment, performing format recognition and conversion on heterogeneous instrument data using a standard protocol adapter, and generating a first data set in a unified format includes: Develop heterogeneous instrument data acquisition middleware to support automatic parsing of mainstream equipment data; The middleware has built-in standard protocol adapters to perform format recognition and conversion for different instrument types and data formats; The middleware regularly obtains data streams based on the data collection frequency and parses them according to the data parsing rules; The parsed data is verified through a data verification mechanism to ensure data integrity; The verified data is converted into a first data set in a unified format according to data conversion logic; The converted data is stored in the specified location and synchronized according to the data synchronization frequency; The middleware uses data security policies to protect data privacy and sets data access permissions to control data access.

3. The method according to claim 1, characterized in that The first data set is automatically parsed using data middleware, and is integrated into a structured second data set based on metadata rules for field mapping and unit conversion, including: Using data middleware to perform automatic parsing, based on metadata rules for field mapping and unit conversion, the first data set is integrated into a structured second data set; If the data format is text, parse the field name and field type; If the data format is binary, the parsing rules are determined based on the device type; Unify data units according to unit types and conversion rules; Associate device status and experimental records through data source identification and timestamp; Write the parsed data into a structured target to complete data integration.

4. The method according to claim 1, wherein The method of constructing a multi-source data fusion model based on the second data set, classifying the data by instrument type and aggregating them by timestamp through batch processing to generate a fused third data set includes: Build a time series database cluster, divide data into shards by instrument type, design a shard key that includes the device fingerprint field, establish an aggregation strategy for minute-level time windows, and configure stream processing jobs to clean and transform the raw data stream.

5. The method according to claim 1, wherein The third data set is checked for data integrity. When the missing rate is lower than a preset threshold and the format is consistent, the dynamic process engine loads the detection process template to generate an initial process configuration. If it is not satisfied, the second data set is returned to supplement the missing data and then re-parsed, including: Checking the data integrity of the third data set to determine whether the missing rate is lower than a preset threshold and the format is consistent; If the conditions are met, the detection process template is loaded through the dynamic process engine to generate the initial process configuration; If not, return to the second data set to supplement the missing data and re-analyze; Use built-in 19 types of quality control models to perform real-time quality control on data, record all data modification traces, and support three-dimensional traceability by personnel, operation, and impact results; Use RFID technology to track hazardous chemicals throughout their entire life cycle, and link inventory thresholds to the procurement system; The temperature and humidity sensor data is written into the test record in real time, and the experiment will be automatically suspended if it exceeds the standard; Establish a multi-dimensional capability matrix to automatically match testing tasks with personnel qualifications; Direct instrument connection via HTTP text protocol, automatic recording of usage frequency and triggering calibration reminders; Built-in standard detection method library, automatic synchronization of experimental templates when version changes; It supports dynamic adjustment of the detection process, realizes nested electronic signatures from entrustment registration to report issuance, and embeds blockchain timestamps at key nodes.

6. The method according to claim 1, characterized in that For the initial process configuration, the visual design tool is used to adjust the node order and parameter thresholds to generate an optimized process configuration that adapts to various scenarios, including: The process node topology diagram is displayed in the visual designer, and the node type and connection relationship are marked. If the node type is instrument operation, the device type and data collection frequency are associated. If the node type is quality control inspection, the quality control rules and exception handling logic are embedded. According to the detection scenario classification, the node sequence and parameter threshold range are adjusted, the personnel qualification requirements and material properties are matched, the environmental condition monitoring points are set, the task priority and report generation rules are configured, and the optimized process configuration plan is generated.

7. The method according to claim 1, characterized in that According to the optimized process configuration, the process scheduling is performed through the dynamic process engine, the execution time and resource usage are monitored in real time, and the process execution log is generated, including: Establish a multi-dimensional capability matrix, automatically match detection tasks with personnel qualifications, realize direct connection of instruments through HTTP protocol text protocol, automatically record usage frequency and trigger calibration reminders, use RFID technology to realize full life cycle tracking of hazardous chemicals, inventory threshold linkage procurement system, built-in standard detection method library, automatic synchronization of experiment templates with version changes, real-time writing of temperature and humidity sensor data into detection records, automatic suspension of experiments when exceeding standards, intelligent task routing engine supports dynamic adjustment of detection process, paperless whole process from entrustment registration to report issuance, realize electronic signature nesting, key nodes are embedded in blockchain timestamps, establish 19 types of quality control models, audit tracking matrix records all data modification trajectories, and supports three-dimensional traceability based on personnel-operation-impact results.

8. The method according to claim 1, characterized in that The process execution log uses a statistical analysis algorithm to detect quality control indicators. When the indicators deviate from the preset threshold, an early warning signal is generated and the process configuration parameters are updated and optimized. If the threshold is met, a test result report is generated, including: A process log analysis engine was established, and a sliding time window algorithm was used to process time series data to calculate task turnover rate and equipment idle rate indicators.

9. The method according to claim 1, characterized in that The method updates the metadata mapping rules of the multi-source data fusion model based on the detection result report, adjusts the field priority and applies it to the second data set, and continuously improves the data integration efficiency, including: Multi-source data fusion technology is used to match the collected test result data with the metadata mapping rule library. If the test result data is inconsistent with the existing rules, the rule update mechanism is triggered, the field priority is adjusted, the metadata mapping rule library is updated, and the updated rules are applied to the second data set. Through the intelligent task routing engine, the updated data set is associated with the test process. Combined with the personnel qualification matrix, the test tasks are automatically assigned, and the task assignment results are compared with the quality control model. If the quality control model detects abnormal data, a real-time quality control dashboard warning is triggered, and the warning information is associated with the audit tracking matrix to record the data modification trajectory. Through the dynamic process engine, the updated data set is synchronized with the test process template to continuously improve data integration efficiency.

Images