Laboratory environment monitoring data management system based on Internet of Things platform
By employing real-time data acquisition and intelligent filtering and retransmission mechanisms, the problem of missing time-series data in biosafety assessments has been solved, enabling efficient data processing and safety assessment workflows and improving the efficiency and accuracy of biosafety assessments.
Patent Information
- Application Number
- CN202511939458.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-22
- Publication Date
- 2026-04-14
AI Technical Summary
In safety assessments of biological laboratories, the lack of time-series data leads to a waste of data processing resources and low efficiency in safety assessments. Existing technologies have failed to effectively utilize the real-time communication capabilities of IoT platforms for data retransmission, resulting in some data remaining idle for extended periods.
The environmental monitoring module collects data in real time, the interactive management module extracts evaluation data periodically, and the data cleaning unit selectively cleans the data according to the retransmission parameters. It combines discrete point filtering algorithms and decision evaluation indicators to select retransmission parameters, triggers retransmission to obtain real data, and processes non-retransmission parameters efficiently according to preset strategies.
This avoids the waste of resources caused by unnecessary waiting, shortens the data processing cycle, improves the timeliness of security assessment and the reliability of data processing, and ensures the timeliness and accuracy of security hazard investigation.
Smart Images
Figure CN121864266A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of information management technology, specifically to a laboratory environmental monitoring data management system based on an Internet of Things (IoT) platform. Background Technology
[0002] In the context of phased safety assessments in biological laboratories, the core data support for the laboratory environment monitoring data management system based on the Internet of Things (IoT) platform originates from the multi-dimensional monitoring parameter time-series data acquired by various intelligent collectors under the IoT platform architecture. These monitoring parameters specifically include key indicators that are strongly correlated with the safety of biological laboratories, such as temperature and humidity, pollutant concentration, and microbial content. The completeness and accuracy of this type of time-series data are not only the basic prerequisite for the IoT platform to realize the transformation of data value, but also directly determine the reliability of the biological laboratory safety assessment results.
[0003] However, in actual data transmission, due to the combined effects of complex environmental factors such as electromagnetic interference in the laboratory, base station signal obstruction, and network link fluctuations, as well as unstable transmission link factors, the time-series data of the above monitoring parameters are prone to data loss, which hinders the orderly conduct of subsequent security assessment work.
[0004] In existing technologies, to address the aforementioned issue of missing time-series data, system platforms generally employ traditional data cleaning methods to fill in missing values, without leveraging the real-time communication capabilities of IoT platforms to implement a solution that triggers data retransmission. However, in the specific application scenario of biological laboratory safety assessment, the safety assessment process has inherent logical constraints. It requires waiting for all time-series data corresponding to all monitoring parameters to complete the missing value cleaning and filling before the overall safety assessment operation process can be initiated. This process constraint directly results in some time-series data that has been cleaned and filled in advance remaining idle for a long time, which not only causes an ineffective waste of data processing resources but also indirectly affects the efficiency of the safety assessment work.
[0005] To address the above problems, this invention proposes a solution. Summary of the Invention
[0006] The purpose of this invention is to provide a laboratory environmental monitoring data management system based on an Internet of Things platform, in order to solve the problems mentioned in the background art.
[0007] This invention provides a laboratory environmental monitoring data management system based on an IoT platform, comprising:
[0008] The environmental monitoring module is used to collect environmental monitoring data of the target laboratory in real time. The environmental monitoring data includes the monitoring values of several monitoring parameters.
[0009] The interactive management module is used to extract all environmental monitoring data stored in the assessment period according to the preset extraction rules at each assessment cycle to obtain the environmental assessment data of the target laboratory in the assessment cycle. The environmental assessment data contains time-series data of several monitoring parameters.
[0010] The safety assessment platform includes a data cleaning unit and a data analysis unit. The data cleaning unit is used to clean the environmental assessment data of the target laboratory in one assessment cycle based on whether there are several retransmission parameters in the data, and select to perform data cleaning according to the first cleaning step or the second cleaning step.
[0011] The environmental assessment data is cleaned according to the first cleaning step, which yields the target laboratory's cleaned assessment data and missing log data for the assessment period. The environmental assessment data is cleaned according to the second cleaning step, which yields only the target laboratory's cleaned assessment data for the assessment period.
[0012] The data analysis unit is used to analyze the target laboratory's missing log data and assessment cleansing data stored within it for several assessment cycles after they reach a preset fixed amount, and to select a number of monitoring parameters from all monitoring parameters as retransmission parameters.
[0013] Furthermore, the extraction rules for the environmental assessment data of the target laboratory during the assessment period are as follows:
[0014] For any monitoring parameter, taking the start time of the evaluation period as the starting point, and based on the preset extraction duration of the monitoring parameter, the end time of the monitoring parameter in the evaluation period is calculated. Then, from all the environmental monitoring data of the target laboratory stored, all monitoring values corresponding to the monitoring parameter that belong to the evaluation period and whose collection time falls within the interval from the start time to the end time are selected and arranged into continuous sequence data in chronological order, which are used as the time series data of the monitoring parameter in the evaluation period, wherein the end time = start time + extraction duration.
[0015] Environmental assessment data for the target laboratory during the assessment period are generated based on time-series data of all monitoring parameters during the assessment period.
[0016] Furthermore, the analysis steps for selecting several monitoring parameters as retransmission parameters from all monitoring parameters are as follows:
[0017] S11: Mark the missing log data of the target laboratory in all past evaluation periods as A1, A2, ..., Aa, where a≥1;
[0018] All monitoring parameters selected by the management personnel of the target laboratory based on the safety assessment requirements of the target laboratory will be labeled as B1, B2, ..., Bb, where b≥1;
[0019] S12: Calculate and obtain the comprehensive cleaning index H1 of monitoring parameter B1 according to the preset index calculation rules. The index calculation rules are as follows:
[0020] S121: Traverse the missing log data A1. If it does not contain any of the completion values of monitoring parameter B1, do not process it. Otherwise, calculate the completion difference and processing time of the missing log data A1 relative to monitoring parameter B1 according to the preset calculation rules.
[0021] S122: According to S121, traverse the missing log data A2, A3, ..., Aa in sequence, and calculate the missing log data relative to monitoring parameter B1 and the average processing time for each missing log data that does not contain the completion value of monitoring parameter B1.
[0022] S123: Based on all the compensation differences and processing times of the relative monitoring parameter B1 obtained by calculation, the compensation difference G1 and processing time G2 of the monitoring parameter B1 are calculated using the discrete point filtering algorithm.
[0023] S124: Using the formula The comprehensive cleaning index H1 is calculated and obtained from monitoring parameter B1. In the formula... , The numbers are, in order, the imputation difference after normalization and the processing time. α1 and α2 are the preset imputation difference weight and processing time weight, respectively. The cleaning comprehensive index is defined by humans to characterize the accuracy of missing value imputation and the efficiency of processing during the data processing of time series data.
[0024] S13: Calculate and obtain the comprehensive cleaning index H2, H3, ..., Hb of monitoring parameters B2, B3, ..., Bb sequentially according to S12;
[0025] S14: The evaluation cleansing data that belong to the same evaluation period as the missing log data A1, A2, ..., Aa are marked as I1, I2, ..., Ia;
[0026] S15: Generate a sorted list of evaluation cleaned data I1 according to the preset generation rules;
[0027] S16: Generate a sorted list of evaluation cleaned data I2, I3, ..., Ia in sequence according to S15;
[0028] S17: Traverse the sorted list of the cleaned evaluation data I1, I2, ..., Ia, extract all the monitoring parameters contained therein, and for any extracted monitoring parameter, calculate the decision evaluation index of the monitoring parameter based on its occurrence frequency, as detailed below:
[0029] Obtain the frequency L1 of the monitoring parameter, and use the formula Calculate the frequency of occurrence of the decision evaluation index N1 for obtaining the monitoring parameters. The sorted list of the normalized and cleaned evaluation data I1, I2, ..., Ia contains the total number of the monitored parameters. The cleaned comprehensive index is the normalized monitoring parameter;
[0030] S18: For any extracted monitoring parameter, after calculating its decision evaluation index, compare the decision evaluation index with P2. If the decision evaluation index is greater than or equal to P2, the monitoring parameter is selected as the retransmission parameter; otherwise, no processing is performed. P2 is a preset retransmission selection index threshold.
[0031] Furthermore, in S121, the calculation rules for the imputation difference and average processing time of the missing log data A1 relative to the monitoring parameter B1 are as follows:
[0032] S21: Obtain all complete values C1, C2, ..., Cc of monitoring parameter B1 from the missing log data A1, where c ≥ 1; sequentially obtain the monitoring values D1, D2, ..., Dc of the complete values C1, C2, ..., Cc from the missing log data A1; sequentially obtain the processing times E1, E2, ..., Ec of the complete values C1, C2, ..., Cc from the missing log data A1.
[0033] S22: Calculate the completion error Fi of the monitoring parameter B1 within the missing log data A1 using the formula Fi=|Ci-Di|, where i=1, 2, ..., c. In the formula, Ci represents each of the completed values C1, C2, ..., Cc, and Di represents each of the monitored values D1, D2, ..., Dc.
[0034] S23: Using a discrete point filtering algorithm, the completion error Fi and processing times E1, E2, ..., Ec are processed to obtain the completion difference and average processing time of the missing log data A1 relative to the monitoring parameter B1.
[0035] Furthermore, in S15, the rules for generating the sorted list of the evaluation clean data I1 are as follows:
[0036] S151: Obtain the closest acquisition time from the time-series cleaned data of monitoring parameters B1, B2, ..., Bb within the evaluation cleaned data I1, and label them as J1, J2, ..., Jb respectively;
[0037] S152: Calculate the cleaning waiting time Kj for monitoring parameters B1, B2, ..., Bb within the evaluation cleaning data I1 using the formula Kj=Jmax-Jj, j=1, 2, ..., b, where Jmax is the closest acquisition time among J1, J2, ..., Jb to the current time;
[0038] S153: Sort the monitoring parameters corresponding to the cleaning time in descending order of cleaning waiting time, and generate a sorted list of evaluation cleaning data I1 based on the first P1 monitoring parameters after sorting, where P1 is a preset screening quantity threshold.
[0039] Furthermore, the cleaning process for the environmental assessment data according to the first cleaning step is as follows:
[0040] For the time series data of any monitoring parameter included in the environmental assessment data, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter to obtain the time series cleaned data and missing record data of the monitoring parameter in the assessment period.
[0041] The generated content of missing record data for any monitoring parameter during the evaluation period is as follows:
[0042] During the data cleaning process for the time-series data of the monitoring parameters in the evaluation period, all identified missing values are obtained and the processing time and imputation value of each missing value are recorded. The processing time of any missing value is the time interval from the moment the missing value is identified to the moment it is processed. The imputation value is the value after the missing value is filled in.
[0043] Based on the acquisition time corresponding to each identified missing value, a missing value request instruction is generated and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value of the corresponding acquisition time to the data cleaning unit.
[0044] The identified completion values, their processing times, and monitoring values are used to generate missing record data for the monitoring parameters during the evaluation period.
[0045] Based on the time-series cleaned data of all the monitoring parameters obtained in the evaluation period, the evaluation cleaned data of the target laboratory in the evaluation period is generated, and based on the missing record data of all the monitoring parameters obtained in the evaluation period, the missing log data of the target laboratory in the evaluation period is generated.
[0046] Furthermore, the cleaning of the environmental assessment data according to the second cleaning step is as follows: For the time series data of any monitoring parameter included in the environmental assessment data, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter. During the data cleaning process, if several missing values are identified, the monitoring parameter is first determined to be a retransmitted parameter. If the monitoring parameter is a retransmitted parameter, a missing value request instruction is generated according to the several missing values and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value at the corresponding collection time to the data cleaning unit, which performs data cleaning on it and obtains the time series cleaned data of the monitoring parameter in the assessment period after the data cleaning is completed.
[0047] If the monitoring parameter is not a retransmission parameter, no processing is performed. The time series data of the monitoring parameter is cleaned according to the cleaning strategy, and the time series cleaned data of the monitoring parameter in the evaluation period is obtained after the data cleaning is completed.
[0048] The assessment cleaning data for the target laboratory in the assessment period is generated based on the time-series cleaning data of all the monitoring parameters obtained in the assessment period.
[0049] Furthermore, the safety assessment platform also includes a safety assessment unit, which is used to input the assessment and cleaning data of the target laboratory in one assessment cycle into the environmental assessment model after receiving the assessment data of the target laboratory in one assessment cycle, and obtain the assessment result data of the target laboratory output by the environmental assessment model in the assessment cycle. The assessment result data includes several safety assessment sub-items, including indicator quantitative scoring, risk level determination and comprehensive assessment coefficient.
[0050] The safety assessment unit will send the assessment results data of the target laboratory during the assessment period to the electronic device of the management personnel of the target laboratory, so that they can view it at any time.
[0051] Compared with existing technologies, it has the following advantages:
[0052] This invention collects environmental monitoring data from the target laboratory in real time through an environmental monitoring module, and periodically extracts environmental assessment data from the target laboratory through an interactive management module. The data cleaning unit selects to clean the environmental assessment data according to the presence or absence of retransmission parameters, and accurately identifies monitoring parameters with long cleaning wait times through a sorting list. For missing values of retransmission parameters, retransmission is triggered during data waiting intervals to obtain accurate data, avoiding resource waste caused by invalid waiting. At the same time, non-retransmission parameters are efficiently processed according to a preset cleaning strategy, significantly shortening the overall data processing cycle, enabling safety assessments to be initiated more quickly, and improving the timeliness of safety hazard investigation.
[0053] This invention analyzes periodically obtained missing log data and evaluation cleaning data through a data analysis unit, combines the imputation difference and processing time to calculate a comprehensive cleaning index, and then determines the decision evaluation index by combining the frequency of occurrence of monitoring parameters in the sorting list. It accurately screens retransmission parameters with high imputation error and long waiting time, and replaces the algorithm estimation with retransmission to obtain the original real data for the missing values of such parameters, which significantly reduces data bias. At the same time, the invention uses a discrete point filtering algorithm to process imputation error and time data, which further improves the reliability of data processing. Attached Figure Description
[0054] Figure 1 This is a system block diagram of the present invention. Detailed Implementation
[0055] 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.
[0056] Please see Figure 1 This application provides a laboratory environmental monitoring data management system based on an Internet of Things platform, including an environmental monitoring module, an interactive management module, and a safety assessment platform;
[0057] The environmental monitoring module is used to collect environmental monitoring data of the target laboratory in real time. The environmental monitoring data includes the monitoring values of several monitoring parameters. The monitoring parameters are selected by the management personnel of the target laboratory based on the safety assessment needs of the target laboratory. These parameters include, but are not limited to, temperature and humidity, concentration of several harmful gases, air pressure, cleanliness, ventilation rate, pressure difference, and other parameters directly related to the safety of the target laboratory. Harmful gases include, but are not limited to, formaldehyde, ammonia, and microbial aerosols. The target laboratory refers to a biological laboratory.
[0058] In this application, the environmental monitoring module consists of several intelligent sensing devices connected to the Internet of Things platform, including but not limited to networked temperature and humidity sensors, wireless gas concentration monitors, and Internet of Things cleanliness detectors.
[0059] The environmental monitoring module transmits the real-time environmental monitoring data of the target laboratory to the interactive management module.
[0060] The interactive management module is used to store the real-time environmental monitoring data of the target laboratory after receiving it.
[0061] The interactive management module is also used to extract environmental assessment data of the target laboratory for each assessment cycle by retrieving all environmental monitoring data stored within the assessment cycle according to a preset extraction rule. The extraction rule is as follows:
[0062] For any monitoring parameter, taking the start time of the evaluation period as the starting point, and based on the preset extraction duration of the monitoring parameter, the end time of the monitoring parameter in the evaluation period is calculated. Then, from all environmental monitoring data of the target laboratory stored, all monitoring values corresponding to the monitoring parameter that belong to the evaluation period and whose collection time falls within the interval from the start time to the end time are selected and arranged into continuous sequence data in chronological order, which are used as the time series data of the monitoring parameter in the evaluation period, wherein the end time = start time + extraction duration.
[0063] Environmental assessment data for the target laboratory during the assessment period are generated based on the time-series data of all monitoring parameters during the assessment period.
[0064] The interactive management module transmits the environmental assessment data of the target laboratory to the safety assessment platform during the assessment cycle.
[0065] The safety assessment platform is used for periodic safety assessments of the target laboratory. The safety assessment platform includes a data cleaning unit, a safety assessment unit, and a data analysis unit.
[0066] The safety assessment platform receives the environmental assessment data of the target laboratory during the assessment period and then transmits it to the data cleaning unit.
[0067] The data cleaning unit pre-stores several preset cleaning strategies, which correspond one-to-one with each monitoring parameter. Each cleaning strategy includes several data cleaning methods and a cleaning weight for each data cleaning method for a certain monitoring parameter. The larger the value of the cleaning weight, the earlier the execution order of the corresponding data cleaning method is. The initial cleaning weight is preset by the administrator based on the characteristics of common problems of each monitoring parameter.
[0068] Data cleaning methods include, but are not limited to, missing value handling, outlier handling, duplicate data removal, data consistency processing, data normalization / standardization, data discretization, class imbalance handling, text data cleaning, data type conversion, and feature engineering.
[0069] After receiving the environmental assessment data of the target laboratory for the assessment period, the data cleaning unit first determines whether it stores several retransmission parameters. If it does not store several retransmission parameters, it cleans the environmental assessment data according to the first cleaning step to obtain the assessment cleaned data and missing log data of the target laboratory for the assessment period, as follows:
[0070] For the time-series data of any monitoring parameter contained therein, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter to obtain the time-series cleaned data and missing record data of the monitoring parameter in the evaluation period;
[0071] The generated content of missing record data for any monitoring parameter during the evaluation period is as follows:
[0072] During the data cleaning process for the time-series data of the monitoring parameters in the evaluation period, all identified missing values are obtained and the processing time and imputation value of each missing value are recorded. The processing time of any missing value is the time interval from the moment the missing value is identified to the moment it is processed. The imputation value is the value after the missing value is filled in.
[0073] Based on the acquisition time corresponding to each identified missing value, a missing value request instruction is generated and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value of the corresponding acquisition time to the data cleaning unit.
[0074] The identified completion values, their processing times, and monitoring values are used to generate missing record data for the monitoring parameters during the evaluation period.
[0075] Based on all the obtained monitoring parameters and the time-series cleaning data of the assessment period, the assessment cleaning data of the target laboratory in the assessment period is generated, and the assessment cleaning data is transmitted to the safety assessment unit.
[0076] Based on the missing record data of all the monitoring parameters obtained in the evaluation period, the missing log data of the target laboratory in the evaluation period is generated, and the missing log data and the evaluation clean data are transmitted together to the data analysis unit for storage.
[0077] If several retransmission parameters are stored, the environmental assessment data is cleaned according to the second cleaning step as follows:
[0078] For the time-series data of any monitoring parameter included in the environmental assessment data, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter. During the data cleaning process, if several missing values are identified, the monitoring parameter is first determined to be a retransmitted parameter. If the monitoring parameter is a retransmitted parameter, a missing value request instruction is generated based on the several missing values and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value at the corresponding acquisition time to the data cleaning unit, which performs data cleaning on it and obtains the time-series cleaned data of the monitoring parameter in the assessment period after the data cleaning is completed.
[0079] If the monitoring parameter is not a retransmission parameter, no processing is performed. The time series data of the monitoring parameter is cleaned according to the cleaning strategy, and the time series cleaned data of the monitoring parameter in the evaluation period is obtained after the data cleaning is completed.
[0080] Based on all the obtained monitoring parameters and the time-series cleaning data of the assessment period, the assessment cleaning data of the target laboratory in the assessment period is generated, and the assessment cleaning data is transmitted to the safety assessment unit.
[0081] The safety assessment unit pre-stores a trained environmental assessment model for conducting a safety assessment of the target laboratory environment.
[0082] After receiving the assessment and cleaning data of the target laboratory in the assessment period, the safety assessment unit inputs it into the environmental assessment model to obtain the assessment result data of the target laboratory in the assessment period output by the environmental assessment model. The assessment result data includes several safety assessment sub-items, including indicator quantitative scoring, risk level determination and comprehensive assessment coefficient.
[0083] The quantitative scoring of the indicators is used to characterize the compliance level of environmental monitoring parameters in biological laboratories, including but not limited to temperature and humidity compliance scores, hazardous gas concentration compliance rate scores, air pressure stability scores, cleanliness level scores, ventilation rate adaptability scores, and pressure difference rationality scores formed based on the monitoring parameters. The quantitative scoring of each indicator is limited to a range of [0 - 100] points, and the scoring results are positively correlated with the compliance level of the indicators. The risk level determination is based on the quantitative scoring of the indicators and the preset risk threshold, and is specifically divided into three levels: low risk, medium risk, and high risk. Each risk level corresponds to a unique scoring threshold range and a clear level determination rule.
[0084] The safety assessment unit will obtain the assessment results data of the target laboratory during the assessment period and send it to the electronic device of the management personnel of the target laboratory, so that they can view it at any time. The electronic device can be any one of mobile phone, laptop and desktop computer.
[0085] The data analysis unit is used to analyze the target laboratory's missing log data and cleaned evaluation data stored within it for several evaluation cycles after they reach a preset fixed amount. The analysis steps are as follows:
[0086] S11: Mark the missing log data of the target laboratory in all past evaluation periods as A1, A2, ..., Aa, where a≥1;
[0087] All monitoring parameters selected by the management personnel of the target laboratory based on the safety assessment requirements of the target laboratory will be labeled as B1, B2, ..., Bb, where b≥1;
[0088] S12: Calculate and obtain the comprehensive cleaning index H1 of monitoring parameter B1 according to the preset index calculation rules. The index calculation rules are as follows:
[0089] S121: Iterate through the missing log data A1. If it does not contain any of the completed values for monitoring parameter B1, do not perform any processing. Otherwise, calculate the completion difference and average processing time of the missing log data A1 relative to monitoring parameter B1 according to the preset calculation rules. The calculation rules are as follows:
[0090] S21: Obtain all complete values C1, C2, ..., Cc of monitoring parameter B1 from the missing log data A1, where c ≥ 1; sequentially obtain the monitoring values D1, D2, ..., Dc of the complete values C1, C2, ..., Cc from the missing log data A1; sequentially obtain the processing times E1, E2, ..., Ec of the complete values C1, C2, ..., Cc from the missing log data A1.
[0091] S22: Calculate the completion error Fi of the monitoring parameter B1 within the missing log data A1 using the formula Fi=|Ci-Di|, where i=1, 2, ..., c. In the formula, Ci represents each of the completed values C1, C2, ..., Cc, and Di represents each of the monitored values D1, D2, ..., Dc.
[0092] S23: Using a discrete point filtering algorithm, the completion error Fi and processing times E1, E2, ..., Ec are processed to obtain the completion difference and average processing time of the missing log data A1 relative to the monitoring parameter B1. The specific details are as follows:
[0093] The completion error Fi is processed using a discrete point filtering algorithm, and the average value of all remaining completion errors after data processing is calculated. The average value is then calibrated as the completion difference of the missing log data A1 relative to the monitoring parameter B1.
[0094] Similarly, the discrete point filtering algorithm is used to process the processing times E1, E2, ..., Ec, and the average value of all processing times after data processing is calculated. The average value is then defined as the average processing time of the missing log data A1 relative to the monitoring parameter B1.
[0095] In this application, the discrete point filtering algorithm is the H-score filtering algorithm;
[0096] S122: According to S121, traverse the missing log data A2, A3, ..., Aa in sequence, and calculate the missing log data relative to monitoring parameter B1 and the average processing time for each missing log data that does not contain the completion value of monitoring parameter B1.
[0097] S123: Based on the calculated relative monitoring parameter B1, all the completed differences and processing times are calculated: The completed difference G1 and processing time G2 of the monitoring parameter B1 are calculated using the discrete point filtering algorithm, as follows:
[0098] The discrete point filtering algorithm is used to process all the completion differences of the obtained relative monitoring parameter B1, and the average value of all remaining completion differences after data processing is calculated. The average value is then calibrated as the completion difference G1 of the monitoring parameter B1.
[0099] Similarly, the discrete point filtering algorithm is used to process the processing time of all the relative monitoring parameters B1, and the average value of all remaining average times after data processing is calculated. The average value is then calibrated as the processing time G2 of monitoring parameter B1.
[0100] S124: Using the formula The comprehensive cleaning index H1 is calculated and obtained from monitoring parameter B1. In the formula... , The following are the normalized imputed difference and the processing time, respectively. The normalization process is used to remove the influence of the dimensions on the calculation. α1 and α2 are the preset imputed difference weight and the processing weight, respectively.
[0101] It should be noted that the comprehensive cleaning index is a human-defined indicator used to characterize the accuracy of missing value imputation and the efficiency of processing during the data processing of time series data, so as to make a trade-off between quality and cost.
[0102] S13: Calculate and obtain the comprehensive cleaning index H2, H3, ..., Hb of monitoring parameters B2, B3, ..., Bb sequentially according to S12;
[0103] S14: The evaluation cleansing data that belong to the same evaluation period as the missing log data A1, A2, ..., Aa are marked as I1, I2, ..., Ia;
[0104] S15: Generate a sorted list of the evaluation and cleaning data I1 according to the preset generation rules, which are as follows:
[0105] S151: Obtain the closest acquisition time from the time-series cleaned data of monitoring parameters B1, B2, ..., Bb within the evaluation cleaned data I1, and label them as J1, J2, ..., Jb respectively;
[0106] S152: Calculate the cleaning waiting time Kj for monitoring parameters B1, B2, ..., Bb within the evaluation cleaning data I1 using the formula Kj=Jmax-Jj, j=1, 2, ..., b, where Jmax is the closest acquisition time among J1, J2, ..., Jb to the current time;
[0107] S153: Sort the monitoring parameters corresponding to the cleaning time in descending order of cleaning waiting time, and generate a sorted list of the first P1 monitoring parameters after sorting to evaluate the cleaning data I1. In the sorted list, the cleaning waiting time of the monitoring parameters decreases sequentially from left to right. P1 is a preset screening quantity threshold.
[0108] S16: Generate a sorted list of evaluation cleaned data I2, I3, ..., Ia in sequence according to S15;
[0109] S17: Traverse the sorted list of the cleaned evaluation data I1, I2, ..., Ia, extract all the monitoring parameters contained therein, and for any extracted monitoring parameter, calculate the decision evaluation index of the monitoring parameter based on its occurrence frequency, as detailed below:
[0110] Obtain the frequency L1 of the monitoring parameter, and use the formula Calculate the frequency of occurrence of the decision evaluation index N1 for obtaining the monitoring parameters. The sorted list of the normalized and cleaned evaluation data I1, I2, ..., Ia contains the total number of the monitored parameters. The cleaned comprehensive index is the normalized monitoring parameter;
[0111] S18: For any extracted monitoring parameter, after calculating its decision evaluation index, compare the decision evaluation index with P2. If the decision evaluation index is greater than or equal to P2, the monitoring parameter is selected as the retransmission parameter; otherwise, no processing is performed. P2 is a preset retransmission selection index threshold.
[0112] The data analysis unit transmits all selected retransmission parameters to the data cleaning unit, which then stores them.
[0113] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.
[0114] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.
Claims
1. A laboratory environmental monitoring data management system based on an Internet of Things platform, characterized in that, include: The environmental monitoring module is used to collect environmental monitoring data of the target laboratory in real time. The environmental monitoring data includes the monitoring values of several monitoring parameters. The interactive management module is used to extract all environmental monitoring data stored in the assessment period according to the preset extraction rules at each assessment cycle to obtain the environmental assessment data of the target laboratory in the assessment cycle. The environmental assessment data contains time-series data of several monitoring parameters. The safety assessment platform includes a data cleaning unit and a data analysis unit. The data cleaning unit is used to clean the environmental assessment data of the target laboratory in one assessment cycle based on whether there are several retransmission parameters in the data, and select to perform data cleaning according to the first cleaning step or the second cleaning step. The environmental assessment data is cleaned according to the first cleaning step, which yields the target laboratory's cleaned assessment data and missing log data for the assessment period. The environmental assessment data is cleaned according to the second cleaning step, which yields only the target laboratory's cleaned assessment data for the assessment period. The data analysis unit is used to analyze the target laboratory's missing log data and assessment cleansing data stored within it for several assessment cycles after they reach a preset fixed amount, and to select a number of monitoring parameters from all monitoring parameters as retransmission parameters.
2. The laboratory environmental monitoring data management system based on an IoT platform according to claim 1, characterized in that, The extraction rules for the environmental assessment data of the target laboratory during the assessment period are as follows: For any monitoring parameter, taking the start time of the evaluation period as the starting point, and based on the preset extraction duration of the monitoring parameter, the end time of the monitoring parameter in the evaluation period is calculated. Subsequently, from all the environmental monitoring data of the target laboratory stored, all monitoring values corresponding to the monitoring parameters that belong to the evaluation period and whose collection time falls within the interval from the start time to the end time are selected and arranged into continuous sequence data in chronological order, which serve as the time series data of the monitoring parameters in the evaluation period, wherein the end time = the start time + the extraction duration; Environmental assessment data for the target laboratory during the assessment period are generated based on time-series data of all monitoring parameters during the assessment period.
3. The laboratory environmental monitoring data management system based on an IoT platform according to claim 1, characterized in that, The analysis steps for selecting several monitoring parameters as retransmission parameters from all monitoring parameters are as follows: S11: Mark the missing log data of the target laboratory in all past evaluation periods as A1, A2, ..., Aa, where a≥1; All monitoring parameters selected by the management personnel of the target laboratory based on the safety assessment requirements of the target laboratory will be labeled as B1, B2, ..., Bb, where b≥1; S12: Calculate and obtain the comprehensive cleaning index H1 of monitoring parameter B1 according to the preset index calculation rules. The index calculation rules are as follows: S121: Traverse the missing log data A1. If it does not contain any of the completion values of monitoring parameter B1, do not process it. Otherwise, calculate the completion difference and processing time of the missing log data A1 relative to monitoring parameter B1 according to the preset calculation rules. S122: According to S121, traverse the missing log data A2, A3, ..., Aa in sequence, and calculate the missing log data relative to monitoring parameter B1 and the average processing time for each missing log data that does not contain the completion value of monitoring parameter B1. S123: Based on all the compensation differences and processing times of the relative monitoring parameter B1 obtained by calculation, the compensation difference G1 and processing time G2 of the monitoring parameter B1 are calculated using the discrete point filtering algorithm. S124: Using the formula The comprehensive cleaning index H1 is calculated and obtained from monitoring parameter B1. In the formula... , The numbers are, in order, the imputation difference after normalization and the processing time. α1 and α2 are the preset imputation difference weight and processing time weight, respectively. The cleaning comprehensive index is defined by humans to characterize the accuracy of missing value imputation and the efficiency of processing during the data processing of time series data. S13: Calculate and obtain the comprehensive cleaning index H2, H3, ..., Hb of monitoring parameters B2, B3, ..., Bb sequentially according to S12; S14: The evaluation cleansing data that belong to the same evaluation period as the missing log data A1, A2, ..., Aa are marked as I1, I2, ..., Ia; S15: Generate a sorted list of evaluation cleaned data I1 according to the preset generation rules; S16: Generate a sorted list of evaluation cleaned data I2, I3, ..., Ia in sequence according to S15; S17: Traverse the sorted list of the cleaned evaluation data I1, I2, ..., Ia, extract all the monitoring parameters contained therein, and for any extracted monitoring parameter, calculate the decision evaluation index of the monitoring parameter based on its occurrence frequency, as detailed below: Obtain the frequency L1 of the monitoring parameter, and use the formula Calculate the frequency of occurrence of the decision evaluation index N1 for obtaining the monitoring parameters. The sorted list of the normalized and cleaned evaluation data I1, I2, ..., Ia contains the total number of the monitored parameters. The cleaned comprehensive index is the normalized monitoring parameter; S18: For any extracted monitoring parameter, after calculating its decision evaluation index, compare the decision evaluation index with P2. If the decision evaluation index is greater than or equal to P2, the monitoring parameter is selected as the retransmission parameter; otherwise, no processing is performed. P2 is a preset retransmission selection index threshold.
4. The laboratory environmental monitoring data management system based on an IoT platform according to claim 3, characterized in that, S121, the calculation rules for the completion difference and average processing time of the missing log data A1 relative to the monitoring parameter B1 are as follows: S21: Obtain all complete values C1, C2, ..., Cc of monitoring parameter B1 from the missing log data A1, where c ≥ 1; sequentially obtain the monitoring values D1, D2, ..., Dc of the complete values C1, C2, ..., Cc from the missing log data A1; sequentially obtain the processing times E1, E2, ..., Ec of the complete values C1, C2, ..., Cc from the missing log data A1. S22: Calculate the completion error Fi of the monitoring parameter B1 within the missing log data A1 using the formula Fi=|Ci-Di|, where i=1, 2, ..., c. In the formula, Ci represents each of the completed values C1, C2, ..., Cc, and Di represents each of the monitored values D1, D2, ..., Dc. S23: Using a discrete point filtering algorithm, the completion error Fi and processing times E1, E2, ..., Ec are processed to obtain the completion difference and average processing time of the missing log data A1 relative to the monitoring parameter B1.
5. A laboratory environmental monitoring data management system based on an IoT platform according to claim 3, characterized in that, S15, the rules for generating the sorted list of the evaluation cleaned data I1 are as follows: S151: Obtain the closest acquisition time from the time-series cleaned data of monitoring parameters B1, B2, ..., Bb within the evaluation cleaned data I1, and label them as J1, J2, ..., Jb respectively; S152: Calculate the cleaning waiting time Kj for monitoring parameters B1, B2, ..., Bb within the evaluation cleaning data I1 using the formula Kj=Jmax-Jj, j=1, 2, ..., b, where Jmax is the closest acquisition time among J1, J2, ..., Jb to the current time; S153: Sort the monitoring parameters corresponding to the cleaning time in descending order of cleaning waiting time, and generate a sorted list of evaluation cleaning data I1 based on the first P1 monitoring parameters after sorting, where P1 is a preset screening quantity threshold.
6. The laboratory environmental monitoring data management system based on an IoT platform according to claim 1, characterized in that, The environmental assessment data is cleaned according to the first cleaning step as follows: For the time series data of any monitoring parameter included in the environmental assessment data, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter to obtain the time series cleaned data and missing record data of the monitoring parameter in the assessment period. The generated content of missing record data for any monitoring parameter during the evaluation period is as follows: During the data cleaning process for the time-series data of the monitoring parameters in the evaluation period, all identified missing values are obtained and the processing time and imputation value of each missing value are recorded. The processing time of any missing value is the time interval from the moment the missing value is identified to the moment it is processed. The imputation value is the value after the missing value is filled in. Based on the acquisition time corresponding to each identified missing value, a missing value request instruction is generated and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value of the corresponding acquisition time to the data cleaning unit. The identified completion values, their processing times, and monitoring values are used to generate missing record data for the monitoring parameters during the evaluation period. Based on the time-series cleaned data of all the monitoring parameters obtained in the evaluation period, the evaluation cleaned data of the target laboratory in the evaluation period is generated, and based on the missing record data of all the monitoring parameters obtained in the evaluation period, the missing log data of the target laboratory in the evaluation period is generated.
7. The laboratory environmental monitoring data management system based on an IoT platform according to claim 1, characterized in that, The second cleaning step for cleaning the environmental assessment data is as follows: For the time-series data of any monitoring parameter included in the environmental assessment data, data cleaning is performed according to the cleaning weight in the cleaning strategy of the monitoring parameter. During the data cleaning process, if several missing values are identified, the monitoring parameter is first determined to be a retransmitted parameter. If the monitoring parameter is a retransmitted parameter, a missing value request instruction is generated based on the several missing values and transmitted to the interaction management module. After receiving the transmitted missing value request instruction, the interaction management module retransmits the monitoring value at the corresponding collection time to the data cleaning unit, which performs data cleaning on it and obtains the time-series cleaned data of the monitoring parameter in the assessment period after the data cleaning is completed. If the monitoring parameter is not a retransmission parameter, no processing is performed. The time series data of the monitoring parameter is cleaned according to the cleaning strategy, and the time series cleaned data of the monitoring parameter in the evaluation period is obtained after the data cleaning is completed. The assessment cleaning data for the target laboratory in the assessment period is generated based on the time-series cleaning data of all the monitoring parameters obtained in the assessment period.
8. The laboratory environmental monitoring data management system based on an IoT platform according to claim 1, characterized in that, The safety assessment platform also includes a safety assessment unit, which is used to input the assessment and cleaning data of the target laboratory in one assessment cycle into the environmental assessment model after receiving the assessment data of the target laboratory in one assessment cycle, and obtain the assessment result data of the target laboratory output by the environmental assessment model in the assessment cycle. The assessment result data includes several safety assessment sub-items, including indicator quantitative scoring, risk level determination and comprehensive assessment coefficient. The safety assessment unit will send the assessment results data of the target laboratory during the assessment period to the electronic device of the management personnel of the target laboratory, so that they can view it at any time.