Field bus online intelligent diagnosis management and control method and related device

Abstract

The application provides a field bus online intelligent diagnosis management and control method and related device, including the following steps: performing feature level fusion and decision level fusion on multi-source heterogeneous data, and generating a device health state evaluation result based on the fused data; generating early warning information and maintenance strategies based on the device health state evaluation result; the application solves the problems of single function and poor expansibility of existing diagnosis tools, supports high concurrency access of more than 1000 devices, and can be widely applied to industrial scenes such as thermal power generation.

Description

Technical Field

[0001] This invention belongs to the field of industrial automation control technology, and in particular relates to a fieldbus online intelligent diagnostic and control method and related devices. Background Technology

[0002] Fieldbus technology, acting as the neural network of industrial automation control systems, has been widely applied in fields such as thermal power generation, petrochemicals, and metallurgical manufacturing, enabling digital communication and distributed control between field devices and the control system. However, existing fieldbus diagnostic technologies mainly focus on communication protocol analysis and physical layer signal detection, lacking deep perception and intelligent analysis capabilities of equipment operating status, and making seamless integration with upper-level management systems even more difficult.

[0003] Existing fieldbus diagnostic tools typically employ single-function portable diagnostic instruments or PC-based offline analysis software. These tools suffer from technical limitations such as insufficient data acquisition capabilities, simplistic analysis algorithms, and poor scalability. Specifically, traditional diagnostic systems employ a centralized architecture. When the number of connected devices exceeds a certain scale, data acquisition and processing capabilities decline sharply, failing to meet the demands of large thermal power plants for simultaneous online monitoring of over 1000 fieldbus devices. Furthermore, existing systems lack a layered decoupling design, with hardware, algorithms, and application software tightly coupled, leading to difficulties in system upgrades and maintenance, and hindering adaptation to rapidly changing industrial field requirements. In addition, existing diagnostic functions are largely limited to post-fault analysis, lacking pre-fault warning and in-process decision support capabilities, and failing to construct a closed-loop management system for the entire equipment lifecycle. This results in frequent unplanned equipment downtime, severely impacting production safety and economic efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide a fieldbus online intelligent diagnostic and control method and related device, which solves the technical problems of existing diagnostic tools having limited functionality, poor scalability, and inability to provide full lifecycle management.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a fieldbus online intelligent diagnostic and control method, comprising the following steps: Feature-level fusion and decision-level fusion are performed on multi-source heterogeneous data, and equipment health status assessment results are generated based on the fused data; Early warning information and maintenance strategies are generated based on the equipment health status assessment results.

[0006] Preferably, the process further includes the following steps before performing feature-level fusion on multi-source heterogeneous data: The received multi-source heterogeneous data is preprocessed sequentially by denoising, time synchronization, and format standardization to form standardized data.

[0007] Preferably, feature-level fusion and decision-level fusion are performed on multi-source heterogeneous data. Specifically, the method is as follows: Wavelet transform or empirical mode decomposition algorithm is used to extract time-frequency features of multi-source data; The DS evidence theory is used to fuse the obtained time-frequency features to obtain fused data. The weights in the DS evidence theory fusion calculation are dynamically adjusted according to the historical credibility of each data source.

[0008] Preferably, the device health status assessment result is generated based on the fused data, including: The fused data is used as input to a deep convolutional neural network to output a probability distribution of fault types. The fused data and fault type probability distribution are used as input to the Long Short-Term Memory network to output the remaining lifetime prediction. The equipment health index is calculated based on the probability distribution of failure types and the predicted value of remaining service life.

[0009] Secondly, the present invention provides a fieldbus online intelligent diagnostic and control system, which adopts a layered distributed architecture, including a hardware layer, a data layer, an algorithm layer and an application layer. The hardware layer includes an intelligent computing module and at least one diagnostic module, used to collect multi-source heterogeneous data from field devices; The data layer is used to perform noise reduction, synchronization, and format standardization on multi-source heterogeneous data to obtain standardized data. The algorithm layer is used to perform feature-level fusion and decision-level fusion on the multi-source heterogeneous data, and generate equipment health status assessment results based on the fused data. The application layer is used to generate early warning information and maintenance strategies based on the equipment health status assessment results.

[0010] Preferably, the diagnostic module includes a vibration signal acquisition module, a temperature monitoring module, a pressure detection module, and a fieldbus protocol analysis module.

[0011] Preferably, the algorithm layer includes a data fusion unit and a neural network diagnostic unit, wherein: The data fusion unit is used to perform feature-level fusion and decision-level fusion on the multi-source heterogeneous data and output the fused data. The neural network diagnostic unit is used to generate equipment health status assessment results based on the fused data.

[0012] Preferably, the application layer includes a fault early warning module and a maintenance decision support module, wherein: The fault early warning module is used to generate graded early warning information based on the equipment health status assessment results and preset thresholds; The maintenance decision support module is used to generate decision recommendations that include maintenance strategies, spare parts requirements, and personnel scheduling based on the hierarchical early warning information.

[0013] Thirdly, the present invention provides an electronic device including a processor and a memory, wherein the memory stores computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the method described thereon.

[0014] Fourthly, the present invention provides a computer program product, the computer program product including computer-executable instructions, which, when executed, implement the method described.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention provides an online intelligent diagnostic and control method for fieldbus. By performing feature-level and decision-level fusion of multi-source heterogeneous data and generating equipment health status assessment results based on the fused data, it then generates early warning information and maintenance strategies. This solves the problem that existing fieldbus diagnostic tools only focus on communication protocols and physical layer detection, lacking in-depth equipment perception and intelligent analysis. It breaks through the limitations of traditional tools with single functions and simple algorithms, achieving a comprehensive and accurate assessment of equipment operating status. It can provide early warning and maintenance decision support, shifting from passive post-event maintenance to proactive predictive maintenance, effectively reducing unplanned equipment downtime. At the same time, it adapts to the needs of multi-source data processing, and with the system's layered distributed architecture, it can support large-scale concurrent access of equipment, improving the comprehensiveness, real-time performance, and effectiveness of diagnostic and control, and providing core methodological support for closed-loop management of the entire equipment lifecycle.

[0016] Furthermore, due to the adoption of a layered distributed architecture design, the hardware layer, data layer, algorithm layer, and application layer are decoupled. This enables independent upgrades and flexible expansion of each functional module of the system. When a new diagnostic algorithm needs to be added, only the software module of the algorithm layer needs to be updated without modifying the hardware layer configuration. When a new fieldbus protocol needs to be connected, only the corresponding diagnostic module needs to be added to the hardware layer without affecting the normal operation of the upper-layer application software, significantly improving the system's maintainability and technology iteration speed. In actual application testing on a 1000MW ultra-supercritical thermal power generating unit, under the conditions of simultaneously connecting 1200 fieldbus devices and a data sampling frequency of 1kHz, the system's data throughput reached 850MB / s, a 165% improvement compared to the 320MB / s of the existing centralized architecture system. Moreover, when a single hardware module fails, the system can automatically switch to a backup module within 50 milliseconds, reducing service interruption time by 92%.

[0017] Furthermore, by integrating a full lifecycle management mechanism that includes fault early warning and maintenance decision support, the system achieves a shift from passive, reactive maintenance to proactive, preventative, and predictive maintenance. Through the early identification of equipment degradation trends by the neural network diagnostic unit and the tiered early warning system of the fault early warning module, sudden equipment failures and unplanned downtime are effectively avoided. In actual operation tests of feedwater pump units in thermal power plants, the system of this invention provided an early warning lead time of 720 hours for early bearing wear failures, while traditional vibration monitoring systems could only issue alarms 48 hours before a failure occurred. This 15-fold increase in warning time allows maintenance personnel ample time to plan repairs and reduces emergency downtime events by 95%. Attached Figure Description

[0018] Figure 1 This is a system architecture diagram according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the hardware layer modular cabinet structure according to an embodiment of the present invention; Figure 3 This is a flowchart of the adaptive sampling frequency adjustment mechanism according to an embodiment of the present invention; Figure 4 This is a flowchart of the multi-source data fusion and neural network diagnostic algorithm according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the neural network diagnostic unit structure according to an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the deployment of a thermal power plant application scenario according to an embodiment of the present invention; Figure 7 This is a flowchart of the data processing of the data communication layer in an embodiment of the present invention. Detailed Implementation

[0019] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0020] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0021] It should also be understood that the term “and / or” as used in this application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0022] As used in this application specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if detected [the described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once detected [the described condition or event]," or "in response to detection [the described condition or event]."

[0023] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0025] Example 1 This embodiment provides a fieldbus online intelligent diagnostic and control method, which includes the following steps: Feature-level fusion and decision-level fusion are performed on multi-source heterogeneous data, and equipment health status assessment results are generated based on the fused data; Early warning information and maintenance strategies are generated based on the equipment health status assessment results.

[0026] Example 2 This embodiment discloses a fieldbus online intelligent diagnostic and control system. The system adopts a layered distributed architecture, decoupling hardware resources, data management, intelligent algorithms and application functions, and realizes full lifecycle management from data acquisition and intelligent analysis to maintenance and control.

[0027] The system comprises a four-layer architecture: hardware layer, data layer, algorithm layer, and application layer. The hardware layer serves as the system's perception foundation and includes an intelligent computing module, at least one diagnostic module, and a network expansion module. The intelligent computing module communicates with the at least one diagnostic module through the network expansion module. The diagnostic module includes at least one fieldbus protocol analysis module, which supports Profibus, Foundation Fieldbus, HART, or CAN bus protocols and is used to collect operating data from field devices.

[0028] The data layer is responsible for the acquisition, storage and preprocessing of multi-source heterogeneous data, including a data preprocessing unit and a data storage unit. The data preprocessing unit is communicatively connected to the intelligent computing module in the hardware layer to receive raw data and perform noise reduction, synchronization and format standardization processing on the raw data. The data storage unit uses a distributed storage architecture to store the preprocessed data.

[0029] The algorithm layer is deployed independently of the data layer and includes a distributed sampling unit, a data fusion unit, and a neural network diagnostic unit. The distributed sampling unit is communicatively connected to the data preprocessing unit to receive real-time multi-source heterogeneous data through a message queue. The data fusion unit is used to perform feature-level fusion and decision-level fusion on the multi-source heterogeneous data. The neural network diagnostic unit is used to generate device health status assessment results based on the fused data.

[0030] The application layer serves as the interaction interface between the system and the user, and includes a health assessment module, a fault early warning module, and a maintenance decision support module. The health assessment module communicates with the algorithm layer to obtain the equipment health status assessment results and generate a visual report. The fault early warning module generates graded early warning information based on the equipment health status assessment results and preset thresholds. The maintenance decision support module generates decision suggestions including maintenance strategies, spare parts requirements, and personnel scheduling based on the graded early warning information.

[0031] Example 3 Based on Embodiment 2, this embodiment discloses a fieldbus online intelligent diagnostic and control system. The intelligent computing module includes a multi-core processor, a large-capacity memory unit, and a high-speed data bus interface, wherein: The multi-core processor adopts an ARM or RISC-V architecture and has a main frequency of no less than 1.8GHz.

[0032] The high-capacity memory unit includes DDR4 memory chips with capacities ranging from 8GB to 32GB.

[0033] The high-speed data bus interface includes a PCIe 3.0 x4 interface and a Gigabit Ethernet interface.

[0034] The multi-core processor communicates with the network expansion module via a gigabit Ethernet interface; The multi-core processor is connected to the memory unit via a PCIe 3.0 x4 interface.

[0035] The diagnostic module includes a vibration signal acquisition module, a temperature monitoring module, a pressure detection module, and a fieldbus protocol analysis module, wherein: The vibration signal acquisition module includes a high-precision data acquisition card with 24-bit resolution and a 102.4kHz sampling rate, providing a 16-channel IEPE accelerometer input interface, and incorporating a programmable anti-aliasing filter and signal conditioning circuitry for data acquisition. The temperature monitoring module includes a thermocouple input interface and a resistance temperature detector (RTD) input interface.

[0036] The pressure detection module includes a 4-20mA current loop input interface.

[0037] The fieldbus protocol analysis module supports Profibus, Foundation Fieldbus, HART, and CAN bus protocols.

[0038] The network expansion module adopts a switched Ethernet architecture or a ring industrial Ethernet architecture, provides at least 24 Gigabit Ethernet ports, supports the IEEE 802.3ad link aggregation protocol, and ensures communication reliability under high concurrent data traffic.

[0039] Example 4 Based on Embodiment 2, this embodiment discloses a fieldbus online intelligent diagnostic and control system. The data preprocessing unit includes a digital filter, a time synchronizer, and a data format converter, wherein: The digital filter uses either a finite impulse response (FIR) filtering algorithm or an infinite impulse response (IIR) filtering algorithm to process the received raw data.

[0040] The time synchronizer uses the IEEE 1588 precision time protocol to achieve microsecond-level time synchronization.

[0041] The data format converter is used to convert data formats from different fieldbus protocols into a unified JSON or XML format.

[0042] The data storage unit adopts an architecture based on the Hadoop Distributed File System (HDFS) or Ceph distributed storage system, supporting data sharding, redundant backup and elastic expansion. The storage capacity can be configured from TB to PB to meet the long-term storage needs of large-scale industrial data.

[0043] Example 5 Based on Embodiment 2, this embodiment discloses a fieldbus online intelligent diagnostic and control system. The distributed sampling unit adopts an adaptive sampling frequency adjustment mechanism. When the equipment operating parameters are in a stable range, a low-frequency sampling mode is used to save storage space and computing resources. When parameter fluctuations or abnormal trends are detected, it automatically switches to a high-frequency sampling mode to capture transient fault characteristics.

[0044] The data fusion unit includes a feature extraction subunit, a fusion calculation subunit, and a weight allocation subunit. The feature extraction subunit uses wavelet transform algorithm or empirical mode decomposition algorithm to extract time-frequency features of multi-source data. The fusion calculation subunit uses DS evidence theory to realize the fusion calculation of multi-source evidence. The weight allocation subunit dynamically adjusts the fusion weights according to the historical credibility of each data source.

[0045] The neural network diagnostic unit includes a CNN subunit and an LSTM subunit, wherein: The CNN subunit is used to output a probability distribution of fault types based on the fused data; The LSTM sub-unit is used to output a predicted value of remaining useful life based on the fused data and the probability distribution of fault types. The health assessment module is used to calculate the equipment health index based on the probability distribution of the fault type and the predicted value of the remaining service life.

[0046] Example 6 Based on Example 2, this example discloses a fieldbus online intelligent diagnostic and control system. The health assessment module adopts a web-based visual interface, which supports real-time data curves, historical trend charts, three-dimensional equipment model display, and health status dashboard display.

[0047] The fault early warning module includes a threshold setting unit, a trend analysis unit, and an early warning generation unit. The threshold setting unit supports two modes: static threshold and dynamic adaptive threshold. The trend analysis unit uses exponential smoothing or ARIMA time series model to predict the parameter change trend. The early warning generation unit generates first-level, second-level, and third-level early warnings based on the severity of the fault.

[0048] The maintenance decision support module includes a knowledge base unit, an inference engine unit, and an optimization algorithm unit, wherein: The knowledge base unit is configured to store device failure cases, maintenance manuals, and historical maintenance records. The inference engine unit is configured to generate maintenance strategies using a rule-based expert system or a case-based reasoning algorithm. The optimization algorithm unit is configured to use a genetic algorithm or a particle swarm optimization algorithm to optimize the maintenance plan and resource allocation scheme.

[0049] Example 7 Based on Example 2, this example discloses a fieldbus online intelligent diagnostic and control system. The system uses an open system interface to communicate with a distributed control system (DCS), a safety instrumented system (SIS), and a manufacturing execution system (MES). The open system interface supports OPC UA, Modbus TCP / IP, and EtherNet / IP protocols, enabling full lifecycle management from data acquisition and intelligent analysis to maintenance and control.

[0050] In this embodiment, the open system interface includes a data mapping layer, a protocol conversion layer, and a security authentication layer. The data mapping layer maps the data model of the application layer to the data model of a DCS, SIS, or MES system to achieve data format unification and interoperability between different systems. The protocol conversion layer implements data packet format conversion for different communication protocols, and the security authentication layer adopts a two-way authentication mechanism based on digital certificates and a data encryption transmission mechanism.

[0051] Example 8 This embodiment discloses a fieldbus online intelligent diagnostic and control system, taking the online monitoring, diagnostic, and control of the steam turbine and its auxiliary equipment group of a 1000MW ultra-supercritical thermal power generating unit as an example for detailed explanation.

[0052] In thermal power plants, the steam turbine, as the core power equipment, directly affects the safety and economy of the unit. Real-time monitoring of parameters such as turbine shaft vibration, bearing temperature, and cylinder expansion is crucial for preventing major equipment accidents. While traditional turbine monitoring and protection systems (TSI) can provide basic vibration and displacement monitoring, they lack in-depth fault diagnosis capabilities and information exchange capabilities with upper-level management systems, failing to meet the needs of modern smart power plants for full lifecycle equipment management.

[0053] In this embodiment, the hardware layer of the fieldbus online intelligent diagnostic and control system is deployed in the turbine electronics room, using a standard 19-inch industrial rack. The rack contains a backplane bus, redundant power modules, and 20 functional slots. The intelligent computing modules are inserted into slots 1 and 2 of the rack, employing a dual-machine hot standby configuration. Each intelligent computing module is equipped with a domestically produced Phytium FT-2000 / 4 processor with a 2.2GHz clock speed, 16GB of DDR4 ECC memory, and a 256GB industrial-grade solid-state drive. It connects to the backplane bus via a PCIe 3.0 x4 interface, providing a data exchange bandwidth of up to 16GB / s. The diagnostic module includes 8 vibration signal acquisition modules, 4 temperature monitoring modules, 2 pressure detection modules, and 2 fieldbus protocol analysis modules, inserted into slots 3 through 18, respectively. The vibration signal acquisition module employs a high-precision data acquisition card with 24-bit resolution and a 102.4kHz sampling rate. Each module provides a 16-channel IEPE accelerometer sensor input interface, with built-in programmable anti-aliasing filters and signal conditioning circuitry, enabling direct connection to vibration sensors mounted on the bearing housings of the high, medium, and low-pressure cylinders of the turbine. The temperature monitoring module supports K-type, S-type thermocouples, and Pt100 RTD inputs, with each module providing 32 input channels for monitoring parameters such as bearing bearing temperature, lubricating oil temperature, and cylinder block metal temperature. The pressure detection module provides 16 channels of 4-20mA current loop input for monitoring process parameters such as lubricating oil pressure and condenser vacuum. The fieldbus protocol analysis module supports Profibus-DP and Foundation Fieldbus H1 protocols for connecting to the turbine digital electro-hydraulic control system (DEH) and the fieldbus instruments of the feedwater pump turbine. The network expansion modules are inserted into slots 19 and 20, with a redundant configuration. Each module provides 24 Gigabit Ethernet optical and electrical ports, supports IEEE 802.3ad link aggregation and RSTP rapid spanning tree protocol, and is connected to the core switch in the central control room via optical fiber to form a reliable industrial Ethernet backbone network.

[0054] In the implementation of the data layer, the data preprocessing unit is deployed in an embedded Linux operating system within the intelligent computing module, employing a multi-threaded parallel processing architecture. After the raw data is acquired from the diagnostic module, it first undergoes noise reduction processing via a digital filter. In this embodiment, an FIR bandpass filter is used for the vibration signal, with a passband frequency set to 10Hz to 10kHz to eliminate low-frequency trend terms and high-frequency electromagnetic interference; an IIR low-pass filter is used for the temperature signal, with a cutoff frequency of 0.1Hz to smooth measurement noise.

[0055] After filtering, the data enters the time synchronizer, which synchronizes with the plant's GPS clock source through the IEEE 1588 Precise Time Protocol (PTP) to achieve microsecond-level time alignment of all acquisition channels and ensure that multi-source data from different diagnostic modules have a unified time reference.

[0056] Subsequently, the data format converter parses and converts data frames from different protocols such as Profibus and FF bus into a unified JSON format. Each data point contains fields such as timestamp, device ID, measurement point name, physical quantity value, and quality flag.

[0057] The processed data is transmitted to the data storage unit via an Apache Kafka message queue. The data storage unit adopts a Ceph-based distributed storage cluster consisting of three storage servers, each configured with 12 8TB enterprise-grade hard drives. Erasure Coding technology is used to achieve data redundancy, with a total available storage capacity of 192TB, capable of storing at least 5 years of high-frequency historical data.

[0058] The data storage unit adopts a tiered storage strategy: hot data (last 7 days) is stored in the SSD cache layer, warm data (7 days to 3 months) is stored in the SAS hard disk layer, and cold data (more than 3 months) is automatically migrated to the SATA hard disk layer to balance storage cost and access performance.

[0059] The implementation of the algorithm layer is the core technical aspect of this embodiment. It is deployed on a dedicated high-performance computing server and connected to the data layer via a 10 Gigabit Ethernet. The distributed sampling unit consumes data from the Kafka message queue in real time and adopts an adaptive sampling strategy: when the turbine is in a stable operating state (speed fluctuation less than 0.5%, vibration intensity less than 4.5 mm / s), the system adopts a low-frequency sampling mode, with the vibration signal sampling interval set to 1 second and the temperature signal sampling interval set to 10 seconds; when the vibration intensity is detected to exceed 4.5 mm / s or the temperature change rate exceeds 5℃ / minute, the system automatically triggers a high-frequency sampling mode, increasing the vibration signal sampling rate to 1000 points per second to capture transient fault characteristics.

[0060] The data fusion unit performs multi-level fusion of data from different sensors: In the feature-level fusion stage, the wavelet packet decomposition algorithm is used to decompose the vibration signal into three levels to extract the energy features of each frequency band, and the empirical mode decomposition (EMD) is used to extract the intrinsic mode function (IMF) components of the temperature signal. In the decision-level fusion stage, the DS evidence theory is used to perform fusion calculations on multi-source evidence such as vibration, temperature, and pressure. The basic probability allocation function of the DS evidence theory is dynamically adjusted according to the historical diagnostic accuracy of each sensor. The weight of vibration evidence is set to 0.5, the weight of temperature evidence is set to 0.3, and the weight of pressure evidence is set to 0.2.

[0061] The neural network diagnostic unit adopts a deep convolutional neural network (CNN) architecture, including an input layer, two convolutional layers, two pooling layers, three fully connected layers, and an output layer. The first convolutional layer contains 32 3×3 convolutional kernels with a stride of 1, using the ReLU activation function; the first pooling layer uses 2×2 max pooling; the second convolutional layer contains 64 convolutional kernels; the second pooling layer also uses max pooling. The fully connected layers contain 128, 64, and 32 neurons respectively, and the output layer contains 5 neurons, corresponding to the probability outputs of normal state, imbalance fault, misalignment fault, bearing fault, and oil film whirl fault, respectively. The neural network is implemented using the TensorFlow framework, with input data being preprocessed vibration spectrum diagrams and temperature trend sequences, and the probability distribution of various faults is output through the Softmax function.

[0062] To predict the remaining useful life (RUL) of turbine bearings, a Long Short-Term Memory (LSTM) network model is adopted. The input layer receives feature data from the past 30 days, the hidden layer contains 128 LSTM units, and the output layer predicts the RUL value for the next 90 days. The root mean square error (RMSE) of the prediction is controlled within 72 hours.

[0063] The application layer implementation includes the specific implementations of the health assessment module, fault early warning module, and maintenance decision support module. The health assessment module is developed based on a B / S architecture, runs on a high-performance web server, and supports access via a browser. This module displays the turbine's health status in real time, using a percentage-based Health Index (HI) for quantitative assessment. The HI value calculation formula is:

[0064] In the formula, For health index, For the first The weighting coefficient of each measurement point Let be the normalized degradation degree of the i-th measurement point, and n be the total number of measurement points participating in the evaluation. The normalized degradation degree... Calculated based on the deviation between the measured value and the standard value:

[0065] In the formula, Let i be the measured value of the i-th measuring point. The nominal value under rated operating conditions. These are alarm limits.

[0066] when When the value is greater than 90, the device status is displayed as green and normal; when Values ​​between 75 and 90 are displayed as a yellow warning; when... A value between 60 and 75 indicates an orange warning; when... When the value is below 60, it is displayed as a red danger.

[0067] The fault early warning module adopts a graded early warning mechanism. The static threshold is set according to the technical specifications provided by the turbine manufacturer. For example, the alarm threshold for the effective value of bearing vibration velocity (RMS) is set to 7.1 mm / s, and the danger threshold is set to 11.2 mm / s. The dynamic adaptive threshold adopts the 3σ principle and is dynamically adjusted according to the statistical characteristics of the equipment's historical operating data.

[0068] The trend analysis unit uses the ARIMA(1,1,1) time series model to predict the trend of parameter changes. When the predicted value will exceed the alarm threshold within the next 72 hours, the system will automatically generate a level-two warning.

[0069] The maintenance decision support module has a built-in knowledge base of steam turbine maintenance experts for thermal power plants, which contains more than 500 fault rules and maintenance cases. It uses a forward reasoning mechanism to match the current fault characteristics with the rules in the knowledge base to generate a maintenance work order that includes maintenance items, required spare parts, estimated working hours and personnel qualification requirements.

[0070] The optimization algorithm unit optimizes the maintenance plan of all units in the plant based on a genetic algorithm. The goal is to minimize maintenance costs and downtime losses through the fitness function. The constraints include personnel number limits, spare parts inventory limits, and unit grid connection time window limits. The optimal maintenance sequence is solved iteratively through selection, crossover, and mutation operations.

[0071] The implementation of the open system interface ensures seamless integration of this embodiment with the power plant's existing control and management systems. Integration with the DCS system is achieved through the OPC UA protocol. The data mapping layer maps the data model of this embodiment (such as health indices and fault codes) to the DCS's AO (analog output) and DO (digital output) point tables, with a mapping period set to 500 milliseconds to ensure that the DCS operator station can view the equipment health status in real time. Integration with the SIS system is achieved through the Modbus TCP / IP protocol, sending key early warning information to the SIS's safety logic controller via hardwiring or communication to participate in interlocking protection logic. Integration with the MES system is achieved through a RESTful API interface, uploading equipment maintenance records and fault diagnosis reports to the MES's asset management module and synchronously updating the equipment's electronic files. The security authentication layer uses X.509 digital certificates for two-way authentication. All data transmission is encrypted using the TLS 1.3 protocol, employing the AES-256-GCM symmetric encryption algorithm and the ECDHE key exchange mechanism to ensure the information security of the industrial control network.

[0072] In the specific deployment and implementation process, a site survey was first conducted to determine the layout of measuring points for the steam turbine and its auxiliary equipment (including feedwater pump turbine, condensate pump, circulating water pump, etc.), with a total of 96 acceleration sensors, 128 temperature measuring points, and 64 pressure measuring points installed. Subsequently, fieldbus network cabling was carried out, using a hybrid network of shielded twisted-pair cables and fiber optics to connect the field smart instruments to the diagnostic module. After the cabinet installation was completed, system power-on testing and communication debugging were performed to ensure that all hardware modules were functioning correctly. Next, the algorithm model was trained and validated. Historical operating data and fault cases from the past three years were used to train the neural network, employing the Adam optimizer with a learning rate of 0.001, a batch size of 32, and 200 training iterations. Ultimately, a fault identification accuracy of 96.5% was achieved on the test set. After the system was officially put into operation, a full lifecycle management process was established, from data collection and intelligent analysis to maintenance execution: When the system detects abnormal vibration of the turbine #3 bearing and generates a level-two warning, the maintenance decision support module automatically generates a maintenance work order that includes replacing the #3 bearing lubricating oil, checking the bearing clearance, and preparing a spare bearing. After being reviewed by the equipment management personnel, the work order is issued to the maintenance team. After the maintenance is completed, the maintenance record is filled back through the mobile terminal, and the system automatically updates the health record and remaining life prediction model of the bearing, forming a complete closed-loop management.

[0073] Through actual operation verification in this embodiment, the system of this embodiment has achieved concurrent access and real-time diagnosis of 1200 fieldbus devices on a 1000MW thermal power generating unit. The overall availability of the system reached 99.95%, the false alarm rate was less than 2%, and the average advance warning of equipment failure reached 15 days, which significantly improved the operational reliability and maintenance economy of the unit, and fully verified the practicality and advancement of the technical solution of this embodiment.

[0074] Example 9 This embodiment also provides a computing device. The computing device includes a bus, a processor, a memory, and a communication interface. The processor, memory, and communication interface communicate with each other via the bus. The computing device can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memory in the computing device.

[0075] A bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, a bus can include a path for transmitting information between various components of a computing device (e.g., memory, processor, communication interfaces).

[0076] The processor may include any one or more of the following: central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU), application specific integrated circuit (ASIC), field-programmable gate array (FPGA), microprocessor (MP), or digital signal processor (DSP).

[0077] Memory can include volatile memory, such as random access memory (RAM). Processors can also include non-volatile memory. volatile memory, such as read-only memory (ROM). ROM (memory only), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0078] The memory stores executable program code, which the processor executes to implement the functions of the aforementioned units, thereby achieving, for example, the method described in Embodiment 1. That is, the memory may store instructions for the methods and functions relating to the computing device in any of the above embodiments.

[0079] The communication interface uses transceiver modules such as, but not limited to, network interface cards and transceivers to enable communication between computing devices and other devices or communication networks.

[0080] Example 10 This embodiment also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, cause the processor to perform the methods and functions of the computing device involved in any of the above embodiments.

[0081] Generally, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software, which can be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are shown and described as block diagrams, flowcharts, or represented using some other illustration, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as, as non-limiting examples, in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0082] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A fieldbus-based online intelligent diagnostic and control method, characterized in that, Includes the following steps: Feature-level fusion and decision-level fusion are performed on multi-source heterogeneous data, and equipment health status assessment results are generated based on the fused data; Early warning information and maintenance strategies are generated based on the equipment health status assessment results.

2. The fieldbus online intelligent diagnostic and control method according to claim 1, characterized in that, Before performing feature-level fusion on multi-source heterogeneous data, the following steps are also required: The received multi-source heterogeneous data is preprocessed sequentially by denoising, time synchronization, and format standardization to form standardized data.

3. The fieldbus online intelligent diagnostic and control method according to claim 1, characterized in that, The specific methods for feature-level fusion and decision-level fusion of multi-source heterogeneous data are as follows: Wavelet transform or empirical mode decomposition algorithm is used to extract time-frequency features of multi-source data; The DS evidence theory is used to fuse the obtained time-frequency features to obtain fused data. The weights in the DS evidence theory fusion calculation are dynamically adjusted according to the historical credibility of each data source.

4. The fieldbus online intelligent diagnostic and control method according to claim 1, characterized in that, Based on the fused data, device health status assessment results are generated, including: The fused data is used as input to a deep convolutional neural network to output a probability distribution of fault types. The fused data and fault type probability distribution are used as input to the Long Short-Term Memory network to output the remaining lifetime prediction. The equipment health index is calculated based on the probability distribution of failure types and the predicted value of remaining service life.

5. A fieldbus-based online intelligent diagnostic and control system, characterized in that, It adopts a layered distributed architecture, including a hardware layer, a data layer, an algorithm layer, and an application layer; The hardware layer includes an intelligent computing module and at least one diagnostic module, used to collect multi-source heterogeneous data from field devices; The data layer is used to perform noise reduction, synchronization, and format standardization on multi-source heterogeneous data to obtain standardized data. The algorithm layer is used to perform feature-level fusion and decision-level fusion on the multi-source heterogeneous data, and generate equipment health status assessment results based on the fused data. The application layer is used to generate early warning information and maintenance strategies based on the equipment health status assessment results.

6. The fieldbus online intelligent diagnostic and control system according to claim 5, characterized in that, The diagnostic module includes a vibration signal acquisition module, a temperature monitoring module, a pressure detection module, and a fieldbus protocol analysis module.

7. The fieldbus online intelligent diagnostic and control system according to claim 5, characterized in that, The algorithm layer includes a data fusion unit and a neural network diagnostic unit, wherein: The data fusion unit is used to perform feature-level fusion and decision-level fusion on the multi-source heterogeneous data and output the fused data. The neural network diagnostic unit is used to generate equipment health status assessment results based on the fused data.

8. The fieldbus online intelligent diagnostic and control system according to claim 5, characterized in that, The application layer includes a fault early warning module and a maintenance decision support module, wherein: The fault early warning module is used to generate graded early warning information based on the equipment health status assessment results and preset thresholds; The maintenance decision support module is used to generate decision recommendations that include maintenance strategies, spare parts requirements, and personnel scheduling based on the hierarchical early warning information.

9. An electronic device, characterized in that, It includes a processor and a memory, the memory storing computer instructions that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1 to 4.

10. A computer program product, characterized in that, The computer program product includes computer-executable instructions that, when executed, implement the method of any one of claims 1 to 4.

Images