CN115101188A - Predictive maintenance system and decision-making assisting method for biochemical analyzer - Google Patents
Predictive maintenance system and decision-making assisting method for biochemical analyzer Download PDFInfo
- Publication number
- CN115101188A CN115101188A CN202211010974.9A CN202211010974A CN115101188A CN 115101188 A CN115101188 A CN 115101188A CN 202211010974 A CN202211010974 A CN 202211010974A CN 115101188 A CN115101188 A CN 115101188A
- Authority
- CN
- China
- Prior art keywords
- data
- instrument
- management
- maintenance
- biochemical analyzer
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012423 maintenance Methods 0.000 title claims abstract description 98
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000003745 diagnosis Methods 0.000 claims abstract description 29
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 27
- 238000000338 in vitro Methods 0.000 claims abstract description 27
- 238000004458 analytical method Methods 0.000 claims abstract description 19
- 238000013135 deep learning Methods 0.000 claims abstract description 15
- 238000012544 monitoring process Methods 0.000 claims abstract description 11
- 238000007726 management method Methods 0.000 claims description 92
- 238000012545 processing Methods 0.000 claims description 24
- 238000009826 distribution Methods 0.000 claims description 21
- 238000013508 migration Methods 0.000 claims description 21
- 238000004140 cleaning Methods 0.000 claims description 20
- 230000005012 migration Effects 0.000 claims description 19
- 239000000047 product Substances 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 230000006870 function Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 9
- 238000012742 biochemical analysis Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000007405 data analysis Methods 0.000 claims description 7
- 230000036541 health Effects 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 238000007872 degassing Methods 0.000 claims description 6
- 238000009499 grossing Methods 0.000 claims description 6
- 239000012502 diagnostic product Substances 0.000 claims description 5
- 238000007619 statistical method Methods 0.000 claims description 5
- 239000003153 chemical reaction reagent Substances 0.000 claims description 4
- 230000007613 environmental effect Effects 0.000 claims description 4
- 239000000463 material Substances 0.000 claims description 4
- 238000004891 communication Methods 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims description 3
- 238000007418 data mining Methods 0.000 claims description 3
- 239000008367 deionised water Substances 0.000 claims description 3
- 229910021641 deionized water Inorganic materials 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 239000012459 cleaning agent Substances 0.000 claims description 2
- 238000001816 cooling Methods 0.000 claims description 2
- 238000003860 storage Methods 0.000 claims description 2
- 238000005259 measurement Methods 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000010801 machine learning Methods 0.000 description 4
- 230000009467 reduction Effects 0.000 description 4
- 238000011161 development Methods 0.000 description 3
- 238000013439 planning Methods 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000002059 diagnostic imaging Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 238000003756 stirring Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 239000000427 antigen Substances 0.000 description 1
- 102000036639 antigens Human genes 0.000 description 1
- 108091007433 antigens Proteins 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003018 immunoassay Methods 0.000 description 1
- 230000036046 immunoreaction Effects 0.000 description 1
- 238000004020 luminiscence type Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003449 preventive effect Effects 0.000 description 1
- 238000005057 refrigeration Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012384 transportation and delivery Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/40—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the management of medical equipment or devices, e.g. scheduling maintenance or upgrades
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
- G16H40/67—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for remote operation
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- Business, Economics & Management (AREA)
- General Health & Medical Sciences (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Databases & Information Systems (AREA)
- Automatic Analysis And Handling Materials Therefor (AREA)
Abstract
The invention relates to the field of remote management and decision assistance of in-vitro diagnosis equipment, in particular to a predictive maintenance system and a decision assistance method for a biochemical analyzer, which realize the digital and intelligent remote monitoring management of the life cycle of a full-automatic biochemical analyzer, carry out big data statistics by an in-vitro diagnosis product management SaaS platform, realize the intelligent analysis of the full-automatic biochemical analyzer by using a deep learning algorithm, realize the predictive maintenance of the instrument, provide an efficient and scientific decision assistance basis for medical instrument enterprises, make an effective predictive maintenance strategy and plan, reduce the loss caused by the post-maintenance of the instrument to the greatest extent, and improve the working efficiency of the instrument management and maintenance of the medical instrument enterprises.
Description
Technical Field
The invention relates to the field of remote management and decision assistance of in-vitro diagnostic equipment, in particular to a predictive maintenance system and a decision assistance method for a biochemical analyzer.
Background
The full-automatic biochemical analyzer is one of indispensable instruments for clinical diagnosis in current medical institutions, adopts a novel labeled immunoassay technology for detecting trace antigens or antibodies, which is established by combining luminescence analysis and immunoreaction, and replaces the complicated operation flow in the traditional manual biochemical analysis in a full-automatic mode, wherein the operation flow comprises a series of operations of sampling, reagent adding, interference removing, mixing and stirring, constant temperature reaction, automatic detection, data processing, cleaning of a reaction cup after test and the like. The application of the full-automatic biochemical analyzer greatly improves the efficiency and the accuracy of biochemical analysis of medical structures and greatly promotes the modernization and digital development process of medical institutions. For the maintenance of the instrument product related to multidisciplinary cross fusion, the requirement of regular maintenance of the instrument and fault diagnosis personnel is high, so that once a high-precision complex medical instrument such as a full-automatic biochemical analyzer fails, the positioning and reason finding of the instrument fault consume manpower and material resources, meanwhile, the maintenance period is long, the cost is high, the detection working process of a hospital is greatly blocked, and the loss which is difficult to estimate is caused.
Meanwhile, the development of the medical internet of things and the value of the medical internet of things are greatly promoted due to the symbiotic growth of machine learning and artificial intelligence, and when a large amount of data stream information generated by the sensor-assisted medical equipment is processed through a machine learning and artificial intelligence algorithm, a feasible conclusion and a scientific decision can be provided more quickly and accurately compared with manual processing. With the rapid development of internet technology, the medical instrument industry is gradually becoming the field of wide application of artificial intelligence technology, and artificial intelligence medical instruments are medical instruments adopting the artificial intelligence technology, and objective medical instrument data used for medical purposes, such as medical image data generated by medical imaging equipment, physiological parameter data generated by medical electronic equipment, in-vitro diagnosis data generated by in-vitro diagnosis equipment and the like, are collected; in special cases, objective data for medical purposes generated by general-purpose equipment, such as the operational status data of a fully automatic biochemical analyzer collected in the present invention, also belong to the data medical instrument data. The model is trained and evaluated through stages of demand analysis, data collection, algorithm design, verification, confirmation and the like, and the artificial intelligent algorithm model is trained through steps of collection of the running state of the instrument, algorithm design and the like, so that the predictive maintenance suitable for the full-automatic biochemical analyzer is finally realized.
Common maintenance approaches for instruments typically include post-mortem, preventative and predictive maintenance. The subsequent maintenance is to perform maintenance after the instrument fails, and the maintenance mode can cause equipment to be in a failure and stop and high maintenance cost; the preventive maintenance mode is to manually make a regular maintenance plan for the instrument, which may result in untimely maintenance of the instrument or excessive maintenance times, and waste of resources.
Chinese patent ZL201980046364.4 discloses a predictive medical device maintenance management in which systems and methods of managing maintenance of a plurality of monitored medical devices are disclosed, the systems and methods including receiving streaming time series medical device data from the plurality of monitored medical devices, analyzing the streaming time series medical device data to determine an operating state of a component of a medical device of the plurality of monitored medical devices, determining a maintenance procedure for the medical device based on the operating state of the component of the medical device.
Chinese patent ZL200410030458.8 discloses a medical device predictive maintenance method and device in which a technique for scheduling maintenance of a device such as a medical imaging system is disclosed that allows for selection of a time-based or usage-based schedule, and if a usage-based schedule is selected, the operational data collected from the system is used as the basis for calculating a maintenance schedule along with baseline usage values for parameters indicative of usage, which may be referenced to criteria for similar devices, and the schedule may be modified accordingly, whether increased or decreased, trends in usage may be adjusted by comparing usage determinations over time, and the schedule may be adjusted accordingly.
However, the above prior art is not accurate enough for detecting the state of the equipment, so it is difficult to implement accurate maintenance prediction for the equipment, and the judgment of the predictive maintenance manner is still mechanical, and there is no intelligent decision-making capability.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a predictive maintenance system and an auxiliary decision method for a biochemical analyzer, which adopt predictive maintenance based on state, adopt a deep migration learning algorithm to predict the health state of equipment according to the running state data of the analyzer, and make a maintenance strategy according to the predicted change result of the health state of the equipment.
In order to achieve the purpose, the invention provides the following technical scheme:
a biochemical analyzer predictive maintenance system comprises a full-automatic biochemical analyzer, wherein the full-automatic biochemical analyzer comprises an analysis part, an operation part and an output part, the analysis part is an analyzer host part and is used for analyzing samples, measuring clinical chemical components of various samples and generating result data, and the operation part completes test application, test and reaction processes and simultaneously records running state data of the full-automatic biochemical analyzer during running;
the system comprises a data acquisition module, a data acquisition module and a data analysis module, wherein the data acquisition module is used for remotely acquiring state data of a full-automatic biochemical analyzer and communicating with an in vitro diagnostic product management (SaaS) platform and consists of an Internet of things card hardware communication device and remote data receiving system software, remote acquisition of the state data of the analyzer is realized through the software and the hardware together, and the acquired data is preprocessed after the data acquisition is finished, so that the quality of the data is further improved, and the precision and the algorithm performance of a subsequent algorithm learning process are improved;
the in-vitro diagnosis product management SaaS platform is used for communicating with the data acquisition module, performing online predictive maintenance and online arrangement of instrument maintenance and maintenance tasks, big data analysis and assistant decision management, and comprises an instrument management module, a work management module, an operation and maintenance management module, a warehouse-out management module, a configuration management module, a user management module, a statistical analysis module and an assistant decision module;
and carrying out digital and intelligent monitoring management on the whole life cycle of the full-automatic biochemical analysis, and carrying out big data statistics and deep learning algorithm analysis through software in an in-vitro diagnosis product management (SaaS) platform to realize predictive maintenance of the full-automatic biochemical analyzer.
Further, the data acquisition module realizes data acquisition by the following steps:
a, performing preliminary processing, namely automatically realizing preliminary cleaning of a data set through program detection of a data receiving system, wherein the preliminary cleaning comprises missing data item processing, noise smoothing processing, outlier identification and inconsistent data cleaning;
b, data sorting, namely performing data sorting operation after the preliminary treatment is finished, converting partial data based on the data after the preliminary treatment, and converting the original data into a form suitable for algorithm learning through data generalization treatment and data normalization treatment;
and c, performing data reduction, namely performing quantity reduction on the original data through a data reduction strategy, and further reducing the data size while approaching or maintaining the quality integrity of the original data.
Further, the data acquisition module acquires status information values and physical signal values of each relevant component of the full-automatic biochemical analyzer host during the operation period, wherein normal working values of each component of the full-automatic biochemical analysis are set as follows: the effective range of the environmental temperature is 15-40 ℃, the effective range of the temperature of the reaction disc is 36.7-37.3 ℃, the effective range of the cooling temperature of the reagent is 1.0-9.0 ℃, the effective range of the temperature of the cleaning water of the whole machine is 21.0-35.0 ℃, the effective range of the temperature of the cleaning water of the cuvette is 31.0-41.0 ℃, the effective range of the temperature of the cleaning agent of the cuvette is 31.0-41.0 ℃, the effective range of the environmental temperature of the inside of the whole machine is 15.0-40.0 ℃, the effective range of the digital voltage of the main control board plus 5V is 4.50V-5.45V, the effective range of the voltage of the refrigerating sheet is 11.00V-13.00V, the effective range of the lamp voltage of the photometer is 11.6V-12.8V, the effective range of the main vacuum pressure is less than-20 kPa, the effective range of the degassing pressure of the ISE is 36 kPa-25 kPa, the effective range of the degassing pressure of the whole machine is 5kPa-45kPa, and the effective range of the resistivity of the deionized water is more than 1.0M omega.
Further, the in vitro diagnosis product management SaaS platform is used for uniformly managing the life cycle of the full-automatic biochemical analyzer, analyzing and storing the acquired data and assisting in decision making,
the instrument management module comprises instrument model management, instrument material accessory management, subsystem management and instrument inventory management submodules, and the subsystem and accessory information of the instrument with the specific model are displayed through a tree structure;
the work management module comprises plan making and task distribution, specifically plan making and task distribution for instrument maintenance and upgrading;
the operation and maintenance management module is used for monitoring the state of the instrument which is taken out of the warehouse, including monitoring the on-line state and the health state information of the instrument, and simultaneously checking the historical off-line record, the historical fault record and the historical operation and maintenance record of the instrument;
the ex-warehouse management module is used for carrying out ex-warehouse recording on instruments, wherein the ex-warehouse recording comprises factory information recording and factory return information recording;
the configuration management module is used for configuring related information of an in-vitro diagnosis product management (SaaS) platform, and comprises configuration of instrument fault types, configuration of message subject information related to the message acquisition module, configuration of task types and configuration of related system parameters;
the user management module realizes the management of the platform administrator for the tenant or the tenant for the system user to set the authority;
the data processing module carries out preprocessing operation on the acquired original data, including data missing item processing, noise smoothing processing, outlier identification and inconsistent data cleaning, and then carries out data conversion, data generalization processing and data specification to finally generate a data set suitable for algorithm learning;
the auxiliary decision-making module is uploaded to a data receiving center of an in-vitro diagnosis product management (SaaS) platform through a network for intelligent analysis and data mining, a new auxiliary decision-making model and a new maintenance strategy are generated, and a predictive maintenance plan is generated according to the maintenance strategy and is issued to a specified instrument manufacturer for instrument maintenance prompt.
An assistant decision-making method for the predictive maintenance of a biochemical analyzer is based on the predictive maintenance system of the biochemical analyzer, and adopts a characteristic-based migration learning method-migration component analysis (TCA) to train a decision-making model, and mainly comprises the following four steps:
1. mapping cross-domain data to common feature space using feature mapping
2. Distance metric feature distribution variance
3. Minimum optimization strategy back propagation result, updating characteristic mapping parameter
4. And the shared classifier trained by the samples in the minimum source domain performs classification prediction on the target domain according to the distribution similarity characteristics.
It is assumed that there is a feature map
The data distribution of the source domain and the target domain after mapping is more approximate, and the data set is marked by the given source domain
And label-free target data set
，
And
representing the edge distributions of the source domain data and the target domain data, respectively, the distance between the two distributions P and Q is calculated as follows:
wherein
Representing the size of a source data set
Representing the size of the target data set, H representing the regenerated nuclear hilbert space,
is a feature vector of the source data machine,
for the feature vector of the target data set, by introducing a kernel function matrix
And condition matrix
The following were used:
the objective function is equivalent to:
wherein λ ≧ 0 is a trade-off parameter, which can be determined empirically, and further decomposes the K matrix as follows:
using matrices
Further division into
Wherein
The distance calculation can be simplified as follows:
the complexity of W is controlled by a regularization term, simplifying the kernel function as follows:
the collected data set of the preprocessed state information of the multi-model in-vitro diagnostic instrument is used as the input of the model, a deep migration model is generated through training of a deep migration learning algorithm, the next collected instrument state data is classified and predicted, and a prediction result is generated and used as a basic basis for making an instrument predictive maintenance plan.
Compared with the prior art, the invention provides a predictive maintenance system and an auxiliary decision method for a biochemical analyzer, which have the following beneficial effects:
1. the invention realizes the digital and intelligent monitoring management of the whole life cycle of the instrument, realizes the predictive maintenance of the full-automatic biochemical analyzer by carrying out big data statistics and deep learning algorithm analysis through the in-vitro diagnosis product management SaaS platform software, provides a scientific and intelligent instrument predictive maintenance list for medical instrument manufacturers, reduces the loss caused by the post-maintenance of the instrument to the maximum extent, provides a high-efficiency and scientific auxiliary decision basis for the medical instrument enterprises, and improves the working efficiency and the decision level of the medical instrument enterprises;
2. the fault diagnosis of the full-automatic biochemical analyzer can be carried out based on intelligent calculation such as machine learning, deep learning and neural network, the health state of the analyzer can be effectively identified, and a predictive maintenance plan is further formulated. For the deep learning theory, the fault characteristics do not need to be extracted manually, the deep learning theory can independently learn the fault characteristics from the collected data set, and the labor cost is further reduced. Although the deep learning theory has great effect in the field of intelligent fault diagnosis at present, a great deal of label data is needed for training a diagnosis model through the deep learning theory, and a great deal of labeled samples are needed for training a traditional neural network to obtain a relatively excellent classification prediction model, so that the application of an intelligent calculation method in the fields of fault diagnosis and predictive maintenance of large-scale high-precision medical instruments is greatly limited. The invention adopts a deep migration learning method to construct the prediction model, and the deep migration learning method can well solve the problem of model training under the condition of small samples on one hand and can apply the prediction knowledge learned from one or more tasks to other related prediction tasks on the other hand, thereby overcoming the training and establishment of the model under the conditions of insufficient labeled samples and changeable training target objects of the same series.
Drawings
FIG. 1 is a schematic block flow diagram of the predictive maintenance of a biochemical analyzer according to the present invention;
FIG. 2 is a flow chart of a method of predictive maintenance data processing of the present invention;
FIG. 3 is a flowchart of the feature-based deep migration learning model of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention is described in detail below with reference to fig. 1-3, and the data processing method and decision-making aid system for predictive maintenance of a full-automatic biochemical analyzer of the invention comprises a full-automatic biochemical analyzer analysis part, an operation part, a data acquisition module and an in-vitro diagnosis product management SaaS platform; wherein: the analyzer is a main body part of the analyzer and is used for analyzing samples, measuring clinical chemical components of various samples and generating result data; the operation part completes the processes of test application, test and reaction, and records the running state data of the full-automatic biochemical analyzer at the running time of the analyzer; the data acquisition module is used for remotely acquiring state data of the full-automatic biochemical analyzer and communicating with the in-vitro diagnosis product management SaaS platform; the in-vitro diagnosis product management SaaS platform is used for communicating with the data acquisition module, performing online predictive maintenance, online arrangement of instrument maintenance and maintenance tasks, big data analysis, auxiliary decision management and the like. The invention realizes the digital and intelligent monitoring management of the whole life cycle of the instrument, realizes the predictive maintenance of the full-automatic biochemical analyzer by carrying out big data statistics and deep learning algorithm analysis through the in-vitro diagnosis product management SaaS platform software, provides a scientific and intelligent instrument predictive maintenance list for medical instrument manufacturers, reduces the loss caused by the post-maintenance of the instrument to the maximum extent, provides an efficient and scientific auxiliary decision basis for the medical instrument enterprises, and improves the working efficiency and the decision level of the medical instrument enterprises.
The invention adopts predictive maintenance based on state and adopts a deep migration learning algorithm to carry out equipment pair according to the running state data of an instrument.
1. Data acquisition and processing method
The data acquisition module is used for remotely acquiring the running state data of the full-automatic biochemical analyzer and communicating with the in-vitro diagnosis product management SaaS platform. The data acquisition module consists of an Internet of things card hardware communication device and remote data receiving system software, and remote acquisition of state data of the required instrument is realized in a mode of mutually matching the software and the hardware.
After the data acquisition is finished, the acquired data is preprocessed, so that the quality of the data is further improved, and the accuracy and the algorithm performance of the subsequent algorithm learning process are improved. Firstly, the preliminary cleaning work of the data set is automatically realized through the program detection of a data receiving system, and the preliminary cleaning work comprises the operations of data missing item processing, noise smoothing processing, outlier identification, inconsistent data cleaning and the like. And secondly, performing data sorting operation after the data is preliminarily cleaned, converting partial data based on the data after preliminary cleaning, and converting the original data into a form suitable for algorithm learning through operations such as data generalization processing, data normalization processing and the like.
Finally, the quantity of the original data is reduced through a data reduction strategy, the data size is further reduced while the quality integrity of the original data is approached or maintained, and a data acquisition processing flow is shown in fig. 2.
The state information values and physical signal values of all relevant parts of the full-automatic biochemical analyzer host computer collected by the data collection module during the operation period include but are not limited to the following table information, including the names of all parts of the instrument to be collected and the effective range of the corresponding instrument during normal operation.
| Numbering | Name(s) | Effective range |
| 1 | Ambient temperature | 15℃-40℃ |
| 2 | Temperature of reaction plate | 36.7℃-37.3℃ |
| 3 | Reagent refrigeration temperature | 1.0℃-9.0℃ |
| 4 | Temperature of cleaning water of whole machine | 21.0℃-35.0℃ |
| 5 | Temperature of cuvette cleaning water | 31.0℃-41.0℃ |
| 6 | Temperature of cuvette cleaner | 31.0℃-41.0℃ |
| 7 | Internal ambient temperature of the whole machine | 15.0℃-40.0℃ |
| 8 | Main control board +5V digital voltage | 4.50V-5.45V |
| 9 | Main control board +3.3V digital voltage | |
| 10 | Voltage of refrigerating sheet | 11.00V-13.00V |
| 11 | Photometer lamp | 11.6V-12.8V |
| 12 | Sample stirring rod rotorSpeed-up device | |
| 13 | Main vacuum pressure | <-20kPa |
| 14 | ISE degassing pressure | -36kPa~-25kPa |
| 15 | Degassing pressure of the whole machine | 5kPa-45kPa |
| 16 | Resistivity of deionized water | >1.0MΩ |
| … | … | … |
2. In vitro diagnosis product management SaaS platform
The in-vitro diagnosis product management SaaS platform is used for uniformly managing the life cycle of the full-automatic biochemical analysis instrument, analyzing and storing acquired data and assisting a decision-making system. The in-vitro diagnosis product management SaaS platform comprises an instrument management unit, an operation and maintenance management unit, a configuration management unit, a work management unit, a delivery management unit, a user management unit, a data receiving center, a statistical analysis unit and an auxiliary decision unit. The data acquisition module that combines through software and hardware can realize the remote data acquisition of full-automatic biochemical analysis appearance, and the data storage that gathers carries out the statistical analysis of all kinds of data to the data receiving center of external diagnostic product management SaaS platform based on the data of gathering, uses the degree of depth migration algorithm to carry out the training of supplementary decision-making early warning model, and each component part of external diagnostic product management SaaS platform introduces as follows:
1. instrument management
The instrument management module is mainly used for information management of registered instruments, and comprises instrument model management, instrument material accessory management, subsystem management and instrument inventory management submodules, wherein the subsystem and accessory information of instruments of specific models can be effectively displayed through a tree structure, and the instrument inventory management records inventory information of specific instrument models and detailed information related to warehousing and ex-warehousing of specific instruments;
2. work management
The work management module is mainly used for planning and distributing tasks by an operation and maintenance manager, wherein the planning and the task distribution comprise planning of plans such as instrument maintenance and upgrading, and the like, so that operations such as online task distribution, technical personnel assignment, task record storage and the like are realized;
3. operation and maintenance management
The operation and maintenance management module is used for monitoring the state of the instrument which is delivered from the warehouse, including monitoring the online state and the health state information of the instrument, checking the historical offline record, the historical fault record and the relevant information of the historical operation and maintenance record of the instrument, and recording the historical information of the relevant state of the instrument in detail;
4. warehouse-out management
The ex-warehouse management module is mainly responsible for the records of the instruments from the in-warehouse to the out-warehouse, including factory information records and factory return information records;
5. configuration management
The configuration management module is used for remotely managing the configuration of the relevant information of the cloud platform system, and comprises the configuration of an instrument fault type, the configuration of the relevant message subject information of the message acquisition module, the configuration of a task type and the configuration of relevant system parameters;
6. user management
The user management module is mainly responsible for the management of authority setting of a platform administrator for a tenant or the tenant for a system user;
7. data receiving center
The data receiving center module is mainly responsible for receiving and storing remotely acquired instrument state information data;
8. data analysis
The data analysis module is mainly responsible for preprocessing collected original data, including data missing item processing, noise smoothing processing, outlier identification, inconsistent data cleaning, data conversion, data generalization processing and data specification generation, and finally generating a data set suitable for algorithm learning;
9. aid decision module
The assistant decision-making module uploads the data to a data receiving center of an instrument remote management platform through a network for intelligent analysis and data mining, generates a new assistant decision-making model and a new maintenance strategy, and generates a predictive maintenance plan according to the maintenance strategy to be issued to a specified instrument manufacturer for instrument maintenance prompt.
3. Predictive maintenance model
The fault diagnosis of the full-automatic biochemical analyzer can be carried out based on intelligent calculation such as machine learning, deep learning and neural network, the health state of the analyzer can be effectively identified, and a predictive maintenance plan is further formulated.
For the deep learning theory, the fault characteristics do not need to be extracted manually, the deep learning theory can independently learn the fault characteristics from the collected data set, and the labor cost is further reduced. Although the deep learning theory has great effect in the field of intelligent fault diagnosis at present, a great deal of label data is needed for training a diagnosis model through the deep learning theory, and a great deal of labeled samples are needed for training a traditional neural network to obtain a relatively excellent classification prediction model, so that the application of an intelligent calculation method in the fields of fault diagnosis and predictive maintenance of large-scale high-precision medical instruments is greatly limited.
The invention adopts a deep migration learning method to construct the prediction model, and the deep migration learning method can well solve the problem of model training under the condition of small samples on one hand, and can apply the prediction knowledge learned from one or more tasks to other related prediction tasks on the other hand, thereby overcoming the training and establishment of the model under the conditions of insufficient labeled samples and changeable training target objects of the same series. The process of the feature-based deep migration learning model is shown in fig. 3.
The invention adopts a characteristic-based migration learning method, namely migration component analysis (TCA), to train a decision model, and the method mainly comprises the following four steps:
1. mapping cross-domain data to common feature space using feature mapping
2. Distance metric feature distribution variance
3. Minimum optimization strategy back propagation result, updating characteristic mapping parameter
4. And the shared classifier trained by the samples in the minimum source domain classifies the target domain according to the distribution similarity characteristics.
Suppose there is a feature map
So that the data distribution of the source domain and the target domain after mapping is closer, and the data set is marked by the given source domain
And label-free target data set
，
And
representing the edge distributions of the source domain data and the target domain data, respectively, the distance between the two distributions P and Q is calculated as follows:
wherein
Representing the size of a source data set
Representing the size of the target data set, H representing the regenerated nuclear hilbert space,
is a feature vector of the source data machine,
for the feature vector of the target data set, by introducing a kernel function matrix
And condition matrix
The following were used:
the objective function is equivalent to:
wherein λ ≧ 0 is a trade-off parameter, which can be determined empirically, to further decompose the K matrix as follows:
using matrices
Further division into
Wherein
The distance calculation can be simplified as follows:
the complexity of W is controlled by a regularization term, simplifying the kernel function as follows:
the collected data set of the preprocessed state information of the multi-model in-vitro diagnostic instrument is used as the input of the model, a deep migration model is generated through training of a deep migration learning algorithm, the next collected instrument state data is classified and predicted, and a prediction result is generated and used as a basic basis for making an instrument predictive maintenance plan.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (5)
1. A predictive maintenance system for a biochemical analyzer, comprising:
the full-automatic biochemical analyzer comprises an analysis part, an operation part and an output part, wherein the analysis part is a main machine part of the analyzer and is used for analyzing samples, measuring clinical chemical components of various samples and generating result data, and the operation part is used for completing test application, test and reaction processes and simultaneously recording running state data of the full-automatic biochemical analyzer during running;
the system comprises a data acquisition module, a data acquisition module and a data analysis module, wherein the data acquisition module is used for remotely acquiring state data of a full-automatic biochemical analyzer and communicating with an in vitro diagnostic product management (SaaS) platform and consists of an Internet of things card hardware communication device and remote data receiving system software, remote acquisition of the state data of the analyzer is realized through the software and the hardware together, and the acquired data is preprocessed after the data acquisition is finished, so that the quality of the data is further improved, and the precision and the algorithm performance of a subsequent algorithm learning process are improved;
the in-vitro diagnosis product management SaaS platform is used for communicating with the data acquisition module, performing online predictive maintenance and online arrangement of instrument maintenance and maintenance tasks, big data analysis and decision-making assistance management, and comprises an instrument management module, a work management module, an operation and maintenance management module, a warehouse-out management module, a configuration management module, a user management module, a statistical analysis module and a decision-making assistance module;
and carrying out digital and intelligent monitoring management on the whole life cycle of the full-automatic biochemical analysis, and carrying out big data statistics and deep learning algorithm analysis through software in an in-vitro diagnosis product management (SaaS) platform to realize predictive maintenance of the full-automatic biochemical analyzer.
2. The biochemical analyzer predictive maintenance system of claim 1, wherein the data acquisition module enables data acquisition by:
a, performing preliminary processing, namely automatically realizing preliminary cleaning of a data set through program detection of a data receiving system, wherein the preliminary cleaning comprises missing data item processing, noise smoothing processing, outlier identification and inconsistent data cleaning;
b, data sorting, namely performing data sorting operation after the preliminary treatment is finished, converting partial data based on the data after the preliminary treatment, and converting the original data into a form suitable for algorithm learning through data generalization treatment and data normalization treatment;
and c, performing data specification, namely performing quantity specification on the original data through a data specification strategy, and further reducing the data scale while approaching or maintaining the quality integrity of the original data.
3. The predictive maintenance system for biochemical analyzers according to claim 2, wherein the status information values and the physical signal values of the relevant components of the fully automatic biochemical analyzer during the operation period of the host computer of the fully automatic biochemical analyzer are collected by the data collection module, and the normal operation values of the components of the fully automatic biochemical analyzer are set as follows: the effective range of the environmental temperature is 15-40 ℃, the effective range of the temperature of the reaction disc is 36.7-37.3 ℃, the effective range of the cooling temperature of the reagent is 1.0-9.0 ℃, the effective range of the temperature of the cleaning water of the whole machine is 21.0-35.0 ℃, the effective range of the temperature of the cleaning water of the cuvette is 31.0-41.0 ℃, the effective range of the temperature of the cleaning agent of the cuvette is 31.0-41.0 ℃, the effective range of the environmental temperature of the inside of the whole machine is 15.0-40.0 ℃, the effective range of the digital voltage of the main control board plus 5V is 4.50V-5.45V, the effective range of the voltage of the refrigerating sheet is 11.00V-13.00V, the effective range of the lamp voltage of the photometer is 11.6V-12.8V, the effective range of the main vacuum pressure is less than-20 kPa, the effective range of the degassing pressure of the ISE is 36 kPa-25 kPa, the effective range of the degassing pressure of the whole machine is 5kPa-45kPa, and the effective range of the resistivity of the deionized water is more than 1.0M omega.
4. The biochemical analyzer predictive maintenance system according to claim 3, wherein the in vitro diagnostic product management (SaaS) platform is used for unified management of life cycle of a fully automated biochemical analyzer, analysis and storage of collected data, and decision assistance,
the instrument management module comprises instrument model management, instrument material accessory management, subsystem management and instrument inventory management submodules and displays the subsystem and accessory information of an instrument with a specific model through a tree structure;
the work management module comprises plan making and task distribution, specifically plan making and task distribution for instrument maintenance and upgrading;
the operation and maintenance management module is used for monitoring the state of the instrument which is taken out of the warehouse, including monitoring the on-line state and the health state information of the instrument, and simultaneously checking the historical off-line record, the historical fault record and the historical operation and maintenance record of the instrument;
the ex-warehouse management module is used for carrying out ex-warehouse recording on instruments, wherein the ex-warehouse recording comprises factory information recording and factory return information recording;
the configuration management module is used for remotely managing the configuration of the relevant information of the cloud platform system, and comprises the configuration of an instrument fault type, the configuration of message subject information relevant to the message acquisition module, the configuration of a task type and the configuration of relevant system parameters;
the user management module realizes the management of the platform administrator for the tenant or the tenant for the system user to set the authority;
the statistical analysis module carries out preprocessing operation on the acquired original data, including data missing item processing, noise smoothing processing, outlier identification and inconsistent data cleaning, and then carries out data conversion, data generalization processing and data specification to finally generate a data set suitable for algorithm learning;
the assistant decision-making module uploads the data to a data receiving center of an instrument remote management platform through a network for intelligent analysis and data mining, generates a new assistant decision-making model and a new maintenance strategy, and generates a predictive maintenance plan according to the maintenance strategy to be issued to a specified instrument manufacturer for instrument maintenance prompt.
5. An assistant decision-making method for biochemical analyzer predictive maintenance, which adopts the biochemical analyzer predictive maintenance system of any one of claims 1-4, characterized in that the training of decision-making model is performed by using a feature-based migration learning method-migration component analysis (TCA), and mainly comprises the following four steps:
step 1, mapping cross-domain data to a public feature space by using feature mapping;
step 2, distance measurement feature distribution difference;
step 3, the minimum optimization strategy reversely propagates the result, and the characteristic mapping parameters are updated;
step 4, the shared classifier of the sample training in the minimum source domain carries out classification prediction on the target domain according to the distribution similar characteristics;
suppose there is a feature map
The data distribution of the source domain and the target domain after mapping is more approximate, and the data set is marked by the given source domain
And label-free target data set
，
And
representing the edge distributions of the source domain data and the target domain data, respectively, the distance between the two distributions P and Q is calculated as follows:
wherein
Representing the size of a source data set
Representing the size of the target data set, H representing the regenerated nuclear hilbert space,
is a feature vector of the source data machine,
for the feature vector of the target data set, by introducing a kernel function matrix
And condition matrix
The following were used:
the objective function is equivalent to:
wherein λ ≧ 0 is a trade-off parameter, which can be determined empirically, and further decomposes the K matrix as follows:
using matrices
Further division into
Wherein
The distance calculation can be simplified as follows:
the complexity of W is controlled by a regularization term, simplifying the kernel function as follows:
the collected data set of the preprocessed state information of the multi-model in-vitro diagnostic instrument is used as the input of the model, a deep migration model is generated through training of a deep migration learning algorithm, the next collected instrument state data is classified and predicted, and a prediction result is generated and used as a basic basis for making an instrument predictive maintenance plan.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211010974.9A CN115101188A (en) | 2022-08-23 | 2022-08-23 | Predictive maintenance system and decision-making assisting method for biochemical analyzer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211010974.9A CN115101188A (en) | 2022-08-23 | 2022-08-23 | Predictive maintenance system and decision-making assisting method for biochemical analyzer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115101188A true CN115101188A (en) | 2022-09-23 |
Family
ID=83300307
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211010974.9A Pending CN115101188A (en) | 2022-08-23 | 2022-08-23 | Predictive maintenance system and decision-making assisting method for biochemical analyzer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115101188A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118427764A (en) * | 2024-07-04 | 2024-08-02 | 苏州百源基因技术有限公司 | Life prediction method and system for digital PCR instrument components |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017193497A1 (en) * | 2016-05-09 | 2017-11-16 | 包磊 | Fusion model-based intellectualized health management server and system, and control method therefor |
| CN107942854A (en) * | 2017-11-20 | 2018-04-20 | 湖南华瑞达生物科技有限公司 | A kind of automatic clinical chemistry analyzer remote device diagnostics strategy |
| CN113885408A (en) * | 2021-11-03 | 2022-01-04 | 江西省智能产业技术创新研究院 | Remote operation and maintenance system |
-
2022
- 2022-08-23 CN CN202211010974.9A patent/CN115101188A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017193497A1 (en) * | 2016-05-09 | 2017-11-16 | 包磊 | Fusion model-based intellectualized health management server and system, and control method therefor |
| CN107942854A (en) * | 2017-11-20 | 2018-04-20 | 湖南华瑞达生物科技有限公司 | A kind of automatic clinical chemistry analyzer remote device diagnostics strategy |
| CN113885408A (en) * | 2021-11-03 | 2022-01-04 | 江西省智能产业技术创新研究院 | Remote operation and maintenance system |
Non-Patent Citations (2)
| Title |
|---|
| LAGOON_LALA: "TCA笔记3:TCA论文阅读", 《HTTPS://BLOG.CSDN.NET/LAGOON_LALA/ARTICLE/DETAILS/120514537》 * |
| 鄞舯暹: "TCA（迁移成分分析）", 《HTTPS://BLOG.CSDN.NET/WEIXIN_44390521/ARTICLE/DETAILS/104336224》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118427764A (en) * | 2024-07-04 | 2024-08-02 | 苏州百源基因技术有限公司 | Life prediction method and system for digital PCR instrument components |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112085261B (en) | Enterprise production status diagnosis method based on cloud fusion and digital twin technology | |
| JP6129028B2 (en) | Energy consumption prediction method for building power equipment | |
| CN110335168B (en) | Method and system for optimizing power utilization information acquisition terminal fault prediction model based on GRU | |
| CN111539355A (en) | Photovoltaic panel foreign matter detection system and detection method based on deep neural network | |
| CN114580291B (en) | LSTM-VAE-based dynamic equipment health condition assessment method | |
| CN110757510A (en) | Method and system for predicting remaining life of robot | |
| WO2021007845A1 (en) | Cloud-edge collaborative forecasting system and method for aluminum oxide production indexes | |
| CN115718472A (en) | Fault scanning and diagnosing method for hydroelectric generating set | |
| CN115101188A (en) | Predictive maintenance system and decision-making assisting method for biochemical analyzer | |
| CN116932523B (en) | Platform for integrating and supervising third party environment detection mechanism | |
| CN116933626A (en) | Data monitoring method and device based on digital twinning | |
| CN113592017B (en) | Deep learning model standardized training method, management system and processing terminal | |
| CN117829318A (en) | Digital twin system based on large model algorithm | |
| CN117271099B (en) | Automatic space data analysis scheduling system and method based on rule base | |
| Wang et al. | Software defect fault intelligent location and identification method based on data mining | |
| CN117520184A (en) | Test system for developing computer software | |
| CN116911811A (en) | Resume collection artificial intelligence big data system | |
| CN116663904A (en) | Intelligent target supervision system and method based on data analysis | |
| CN116881798A (en) | Conditional gracile causal analysis method based on variable selection and reverse time lag feature selection for complex systems such as weather | |
| CN116127324A (en) | Fault detection method and device for key components in uninterruptible power supply | |
| CN115146837A (en) | GCN-GRU-based order residual completion period prediction method | |
| CN110555067A (en) | method for collecting, processing and integrally managing assembly quality data | |
| CN111965995B (en) | Three-in-one aircraft avionics system health assessment and prediction system and method thereof | |
| CN115169658A (en) | Inventory consumption prediction method, system and storage medium based on NPL and knowledge graph | |
| CN114818460A (en) | Laboratory equipment residual service life prediction method based on automatic machine learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |