Denitration catalyst performance index intelligent query system based on dynamic data fusion and query method thereof

Abstract

The invention discloses a denitration catalyst performance index intelligent query system based on dynamic data fusion and a query method thereof.The system comprises a data acquisition module, a dynamic data fusion module, a performance prediction module and an intelligent query module, and the dynamic data fusion module is electrically connected with the data acquisition module; the dynamic data fusion module performs weight distribution, noise filtering and standardization processing on the data acquired by the data acquisition module based on an entropy weight method and a Kalman filtering algorithm to generate a fusion data set; the performance prediction module is provided with a performance prediction model and used for receiving the generated fusion data set generated by the dynamic data fusion module, and the performance prediction model is used for predicting and outputting service life prediction value data, the denitration efficiency attenuation trend and the poisoning risk level of the catalyst. According to the denitration catalyst performance index intelligent query system based on dynamic data fusion, an intelligent and refined catalyst management scheme can be provided, and the catalyst management efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of environmental protection technology, and in particular to a denitration catalyst performance index intelligent query system and catalyst performance index intelligent query method based on dynamic data fusion. Background Art

[0002] High-pollution and high-emission areas (such as coal-fired power plants and the chemical industry) place higher demands on the performance management of SCR denitrification catalysts. However, traditional management methods mainly rely on laboratory offline testing and empirical management, resulting in pain points such as high data lag, poor integration, sluggish interaction, and serious waste of resources (premature catalyst replacement). Secondly, the query system with a single data source is difficult to adapt to complex working conditions and lacks dynamic data fusion and intelligent prediction capabilities, making it difficult to meet real-time and precise management needs.

[0003] Therefore, it is necessary to build a catalyst management system that is real-time, high-precision and intelligently interactive, which can provide a set of intelligent and refined catalyst management solutions for application scenarios such as coal-fired power plants, steel metallurgy, and chemical industry, significantly improve catalyst management efficiency, and reduce catalyst operation and maintenance costs. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, one purpose of the present invention is to propose a denitrification catalyst performance index intelligent query system based on dynamic data fusion. By making the dynamic data fusion module include a Kalman filter unit and an entropy weight method weight unit, the system can be equipped with dynamic data fusion capabilities; by making the system include a performance prediction module, the system can be equipped with intelligent prediction capabilities to meet the needs of real-time and precise management; by making the interactive interface capable of multi-condition screening and querying catalyst performance indicators, the system's ability to cope with complex working conditions is improved. The system integrates catalyst operation data, laboratory test data and environmental parameters, and constructs a multi-source database covering a certain span of time through dynamic fusion of multi-source heterogeneous data and adaptive prediction technology. It can provide a set of intelligent and refined catalyst management solutions for application scenarios such as coal-fired power plants, steel metallurgy, and chemical industry, significantly improving catalyst management efficiency.

[0005] The present invention also proposes a catalyst performance index intelligent query method based on the above-mentioned denitration catalyst performance index intelligent query system.

[0006] According to an embodiment of the present invention, a denitrification catalyst performance index intelligent query system based on dynamic data fusion includes: a data acquisition module, which is used to obtain real-time operating parameters of the catalyst, storage data of the laboratory database and environmental parameters; a dynamic data fusion module, which is electrically connected to the data acquisition module and is used to process the data acquired by the data acquisition module, the dynamic data fusion module includes a Kalman filter unit and an entropy weight method weight unit, and the dynamic data fusion module performs weight assignment, noise filtering and standardization on the data acquired by the data acquisition module based on the entropy weight method and the Kalman filter algorithm to generate a fused data set; a performance prediction module, the performance prediction module has a performance prediction model and is used to receive the generated fused data set generated by the dynamic data fusion module, the performance prediction model is used to predict and output the life prediction value data of the catalyst, the denitrification efficiency attenuation trend and the poisoning risk level; an intelligent query module, the intelligent query module includes an interactive interface, the interactive interface is used to query the catalyst performance indicators through natural language input and multi-condition screening, and display them in a visual interface.

[0007] According to the intelligent query system for denitrification catalyst performance indicators based on dynamic data fusion according to an embodiment of the present invention, the dynamic data fusion module includes a Kalman filter unit and an entropy weight method weight unit, so that the system can have dynamic data fusion capabilities; by including a performance prediction module, the system can have intelligent prediction capabilities and meet the needs of real-time and precise management; by enabling the interactive interface to filter and query catalyst performance indicators under multiple conditions, the system's ability to cope with complex working conditions is improved. The system integrates catalyst operation data, laboratory test data and environmental parameters, and constructs a multi-source database covering a certain time span through dynamic fusion of multi-source heterogeneous data and adaptive prediction technology. It can provide a set of intelligent and refined catalyst management solutions for application scenarios such as coal-fired power plants, steel metallurgy, and chemical industry, significantly improving catalyst management efficiency.

[0008] According to some embodiments of the present invention, the data acquisition module is used to be electrically connected to a distributed control system to collect real-time operating parameters of the catalyst in the distributed control system in real time, and the real-time operating parameters of the catalyst include temperature, pressure, NOx concentration, and ammonia escape rate; the data acquisition module is used to be electrically connected to a laboratory database to obtain storage data of the laboratory database, and the storage data of the laboratory database includes specific surface area, chemical composition, trace elements and activity; the data acquisition module is used to be electrically connected to an environmental monitoring device to obtain environmental parameters monitored by the environmental monitoring device, and the environmental monitoring device includes a flue gas analyzer and a humidity sensor to monitor SO2 concentration, humidity and oxygen content, and the environmental parameters include SO2 concentration, humidity and oxygen content.

[0009] According to some embodiments of the present invention, an API interface is included, and the denitrification catalyst performance index intelligent query system is integrated with an external control system through the API interface. The API interface includes a data input interface and a data output interface. The data input interface is used to receive real-time operating condition data from a distributed control system, and the data output interface is used to push a catalyst replacement warning signal to the manufacturing execution system; and / or, the dynamic data fusion module has a function of supporting offline-real-time data synchronization and includes offline data loading and real-time data stream processing, the offline data loading includes batch importing historical laboratory test reports and extracting structured data through optical character recognition technology, and the real-time data stream processing includes using Apache Kafka to build a message queue to achieve low-latency transmission and buffering of sensor data, and the low-latency delay time is less than 1 second.

[0010] According to some embodiments of the present invention, the entropy weight method weight unit is used to calculate the weight coefficient based on the reliability, update frequency and correlation of the data source with the catalyst performance. The formula for calculating the weight coefficient is: Among them, V a is the entropy value of the a-th data, p ij is the proportion of standardized data; the Kalman filter unit is used to construct a state space model based on the data collected by the data acquisition module, and eliminate time series noise through a prediction-correction cycle. The state equation and observation equation are: k =Ax k-1 +w k-1 , z k =Hx k +v k , where x k is the system state vector, Z k is the observed value, W k 、V k are process noise and observation noise.

[0011] According to some embodiments of the present invention, the performance prediction model includes an LSTM network structure, which includes: an input layer, two LSTM hidden layers and an output layer, the input layer is used to fuse data feature dimensions, the number of neurons in the two LSTM hidden layers is 64 / 32, and the output data of the output layer includes the remaining life of the catalyst, denitrification efficiency and degree of catalyst deactivation; wherein, the performance prediction model is configured to adopt an incremental learning strategy to dynamically adjust the weights in the entropy weight method weight unit.

[0012] According to some embodiments of the present invention, the intelligent query module includes: a natural language processing unit, which is used to parse the query statement input by the user; a multi-condition filter, which is provided with a menu, which includes a query solution, which includes the number of reactor layers, temperature range, operating time, and NOx removal efficiency threshold. The number of reactor layers is provided with a submenu, which is used to select the catalyst model; a visualization engine, which is used to generate dynamic charts, which include a line chart superimposed with the historical efficiency trend and the prediction curve, a thermal map of the three-dimensional relationship between temperature-SO2 concentration-efficiency, and a maintenance suggestion pop-up window, and the triggering condition of the maintenance suggestion pop-up window is that the predicted life is <15 days or the efficiency decline rate is >5% / month.

[0013] According to some embodiments of the present invention, a blockchain evidence storage unit is included, which includes data on the chain and a smart contract. The data on the chain is used to write at least part of the prediction results and maintenance operation records into a private chain node, and the smart contract is used to generate an unalterable work order and assign it to relevant personnel when a preset condition is triggered.

[0014] The catalyst performance index intelligent query method of the denitration catalyst performance index intelligent query system according to the second embodiment of the present invention includes:

[0015] S1. Multi-source data acquisition: The data acquisition module acquires the real-time operating parameters of the catalyst, the stored data in the laboratory database, and the environmental parameters;

[0016] S2. Dynamic data fusion: The dynamic data fusion module cleans, assigns weights, and filters noise on the data acquired by the data acquisition module to generate the fused data set;

[0017] S3. Performance prediction: inputting the fused data set into the performance prediction model to calculate the catalyst life, efficiency and risk level;

[0018] S4. Result feedback: returning the query result through the interactive interface.

[0019] According to the catalyst performance index intelligent query method of the embodiment of the present invention, by including dynamic data fusion and performance prediction, the system can meet the needs of real-time and precise management, enhance the system's ability to cope with complex working conditions, and significantly improve catalyst management efficiency.

[0020] According to some embodiments of the present invention, the data acquired by the data acquisition module is cleaned, weighted, and noise filtered by the dynamic data fusion module to generate the fused data set, including: data cleaning: eliminating values ​​that exceed the physical range in the data acquired by the data acquisition module; feature extraction: selecting features that are strongly correlated with catalyst performance, and features that are strongly correlated with catalyst performance include temperature mean, SO2 concentration peak, and active component decay rate; weight optimization: recalculating the entropy weight method weights every 24 hours to adapt to changes in data source quality.

[0021] According to some embodiments of the present invention, the output results of the performance prediction model are further stored through blockchain. The output results of the performance prediction model are stored through blockchain, including: data hashing: generating a unique hash value for the prediction result and writing it into the blockchain distributed ledger; and evidence verification: verifying whether the historical prediction records have been tampered with through a query interface.

[0022] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments with reference to the accompanying drawings, in which:

[0024] Figure 1 is a schematic diagram of a denitration catalyst performance index intelligent query system using dynamic data fusion according to some embodiments of the present invention;

[0025] Figure 2 yes Figure 1 Schematic diagram of the intelligent query interface of the intelligent query system in [1].

[0026] Figure 3 yes Figure 1 Flowchart of the dynamic data fusion algorithm in . DETAILED DESCRIPTION

[0027] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention and are not to be construed as limiting the present invention.

[0028] Reference below Figure 1-Figure 3 The present invention describes an intelligent query system for denitration catalyst performance indicators based on dynamic data fusion according to an embodiment of the present invention.

[0029] The denitration catalyst performance index intelligent query system based on dynamic data fusion according to an embodiment of the present invention includes: a data acquisition module, a dynamic data fusion module, a performance prediction module and an intelligent query module.

[0030] The data acquisition module is used to obtain real-time catalyst operating parameters, laboratory database storage data, and environmental parameters. The data acquisition module provides multi-dimensional heterogeneous data, providing a sufficient data source for the subsequent dynamic data fusion module.

[0031] The dynamic data fusion module is electrically connected to the data acquisition module and is used to process the data acquired by the data acquisition module. The dynamic data fusion module includes a Kalman filter unit and an entropy weighting unit. Based on the entropy weighting method and Kalman filter algorithm, the dynamic data fusion module assigns weights, filters noise, and standardizes the data acquired by the data acquisition module to generate a fused data set. The dynamic data fusion module integrates real-time data collected by the data acquisition module, historical laboratory data, and environmental parameters, and possesses dynamic data fusion capabilities. By including a Kalman filter unit and an entropy weighting unit in the dynamic data fusion module, weights are assigned using the entropy weighting method, and noise is eliminated using the Kalman filter. This improves data reliability and enables relevant personnel to more accurately judge the performance of the denitration catalyst.

[0032] The performance prediction module includes a performance prediction model and is used to receive the generated fusion data set generated by the dynamic data fusion module. The performance prediction model is used to predict and output the catalyst's lifespan prediction value, denitrification efficiency decay trend, and poisoning risk level. By including the performance prediction module in the system, the system has intelligent prediction capabilities to meet the needs of real-time and precise management.

[0033] The intelligent query module includes an interactive interface for querying catalyst performance indicators through natural language input and multi-condition screening, and displays them in a visual interface. For example, the interactive interface can include a user input box for entering natural language queries. By adapting the interactive interface to query catalyst performance indicators through natural language input and multi-condition screening, the system's ability to cope with complex operating conditions can be enhanced. This provides an intelligent and refined catalyst management solution for application scenarios such as coal-fired power plants, steel and metallurgy, and the chemical industry, significantly improving catalyst management efficiency.

[0034] According to the intelligent query system for denitrification catalyst performance indicators based on dynamic data fusion according to an embodiment of the present invention, the dynamic data fusion module includes a Kalman filter unit and an entropy weight method weight unit, so that the system can have dynamic data fusion capabilities; by including a performance prediction module, the system can have intelligent prediction capabilities and meet the needs of real-time and precise management; by enabling the interactive interface to filter and query catalyst performance indicators under multiple conditions, the system's ability to cope with complex working conditions is improved. The system integrates catalyst operation data, laboratory test data and environmental parameters, and constructs a multi-source database covering a certain time span through dynamic fusion of multi-source heterogeneous data and adaptive prediction technology. It can provide a set of intelligent and refined catalyst management solutions for application scenarios such as coal-fired power plants, steel metallurgy, and chemical industry, significantly improving catalyst management efficiency.

[0035] According to some embodiments of the present invention, referring to Figure 1 The data acquisition module is used to be electrically connected to the distributed control system to collect the real-time operating parameters of the catalyst in the distributed control system. The real-time operating parameters of the catalyst include temperature, pressure, NOx concentration, and ammonia slip rate. For example, high-precision sensors (such as thermocouples, infrared NOx analyzers, and laser ammonia slip monitors) are deployed at the inlet, outlet, and key points of the SCR reactor, and the sampling frequency is set to 1 time / 2 hours to accurately measure and sample the real-time operating parameters. The data is transmitted to the edge gateway via the Modbus TCP protocol and stored in the distributed control system. The data acquisition module is electrically connected to the distributed control system to sample the real-time operating parameters of the catalyst from the distributed control system.

[0036] According to some embodiments of the present invention, referring to Figure 1-Figure 3 The data acquisition module is used to electrically connect to a laboratory database to retrieve stored data, including catalyst surface area, chemical composition, trace elements, and activity. For example, a surface area analyzer and an X-ray fluorescence spectrometer are deployed within the laboratory. These data are automatically uploaded to a laboratory database (MySQL) and synchronized to a central server every two weeks. By electrically connecting the data acquisition module to the laboratory database and collecting laboratory catalyst test data, the data source is more comprehensive and accurate.

[0037] According to some embodiments of the present invention, referring to Figure 1-Figure 3The data acquisition module is electrically connected to the environmental monitoring equipment to obtain the environmental parameters monitored by the environmental monitoring equipment. The environmental monitoring equipment includes a flue gas analyzer and a humidity sensor to monitor SO2 concentration, humidity, and oxygen content. By electrically connecting the data acquisition module to the environmental monitoring equipment, the data acquisition module can collect the environmental parameters monitored by the environmental monitoring equipment, incorporating the impact of these environmental parameters on catalyst performance into the reference, thereby making the subsequent fitting model more accurate.

[0038] For example, the measurement range of SO2 concentration is 0-1000ppm; the measurement range of humidity is 0-100%RH; and the measurement range of oxygen content is 0-25%.

[0039] For example, deploying a data preprocessing program (Python) on the edge computing node to filter outliers in real time (such as triggering an alarm and eliminating data when the temperature is >420°C)

[0040] According to some embodiments of the present invention, referring to Figure 1-Figure 3 The denitrification catalyst performance index intelligent query system based on dynamic data fusion includes an API interface. The denitrification catalyst performance index intelligent query system is integrated with the external control system through the API interface. The API interface includes a data input interface and a data output interface. The data input interface is used to receive real-time operating data from the distributed control system, and the data output interface is used to push catalyst replacement warning signals to the manufacturing execution system. For example, the input interface of the API interface receives real-time operating data from the DCS system, and the output interface pushes warning signals to the system. By making the denitrification catalyst performance index intelligent query system based on dynamic data fusion include an API interface and supporting the API interface to connect with third-party systems, it can be applied to different industrial scenarios, thereby improving the versatility of the denitrification catalyst performance index intelligent query system based on dynamic data fusion.

[0041] For example, the intelligent query system for denitrification catalyst performance indicators based on dynamic data fusion also includes permission management capabilities, with roles classified into: Administrator (full permissions), Engineer (data query and model parameter adjustment), and Operator (results viewing only). Administrators can adjust model parameters and import laboratory data; engineers can view prediction results and set alarm thresholds; and operators can only access the visualization panel. The system uses an encryption algorithm to encrypt sensor data and query results during transmission, and also records user queries, model updates, and maintenance operations. The integrity of the log files is ensured by the encryption algorithm.

[0042] The dynamic data fusion module supports offline-to-real-time data synchronization and includes offline data loading and real-time data stream processing. Offline data loading includes batch importing of historical laboratory test reports and extracting structured data through optical character recognition technology. Real-time data stream processing uses Apache Kafka to build a message queue for low-latency transmission and buffering of sensor data, with a low latency of less than 1 second. Because laboratory data may be confidential, laboratory data in related technologies is offline. By enabling the dynamic data fusion module to support offline-to-real-time data synchronization and include offline data loading and real-time data stream processing, it is easier for relevant personnel to import offline experimental data into the online process.

[0043] According to some embodiments of the present invention, referring to Figure 1-Figure 3 The entropy weight method weight unit is used to calculate the weight coefficient according to the reliability, update frequency and correlation of the data source with the catalyst performance. The formula for calculating the weight coefficient is: Among them, W a is the weight coefficient, V a is the entropy value of the a-th data, p ij is the proportion of normalized data. By using the entropy weight method to calculate the weight coefficient, the reliability of the weighted data can be made stronger and the output data more accurate.

[0044] The Kalman filter unit is used to construct a state space model based on the data collected by the data acquisition module and eliminate the time series noise through the prediction-correction cycle. The state equation and observation equation are: k =Ax k-1 +w k-1 , z k =Hx k +v k , where x k is the system state vector, Z k is the observed value, W k 、V k is the process noise and observation noise, process noise W k The value range is 0~Q, and the observation noise V k The value range is 0 ~ R. By using the Kalman filter unit to process the data, the environmental noise can be eliminated, the output data can be made more accurate, and the judgment made based on the output data can be more accurate and reliable.

[0045] For example, before using the Kalman filter unit to perform a prediction-correction cycle to eliminate time series noise, the parameters are first initialized. The parameter initialization process includes: first setting the process noise covariance Q = 0.1 and the observation noise covariance R = 1; through the prediction-correction cycle, the smoothed sensor data is output.

[0046] For example, the dynamic data fusion module performs weighted summation to generate a fused data set. The weighted summation formula includes: fused =w1X lab +w2X sensor +w3X env .

[0047] For example, the weight coefficient of laboratory data can be set to W1=0.5, the weight coefficient of data from the distributed control system can be set to W2=0.3, and the weight coefficient of data from the environmental data can be set to W3=0.2; and W1+W2+W3=1 can be ensured. The weights are recalculated based on new data every 24 hours and the weight coefficients are adjusted dynamically.

[0048] For example, W1, W2, and W3 can be made to be based on The value calculated by the formula.

[0049] For example, the data generated by the dynamic data fusion module can be stored in a time series database (InfluxDB) for model call.

[0050] According to some embodiments of the present invention, referring to Figure 1-Figure 3 The performance prediction model includes an LSTM network structure, which includes: an input layer, two LSTM hidden layers and an output layer. The input layer is used to fuse data feature dimensions. The number of neurons in the two LSTM hidden layers is 64 / 32. The output data of the output layer includes the remaining life of the catalyst, the denitrification efficiency and the degree of catalyst deactivation. Among them, the performance prediction model is configured to adopt an incremental learning strategy to dynamically adjust the weights in the entropy weight method weight unit.

[0051] For example, the input layer of the LSTM network structure includes an N×N feature matrix composed of multi-source data, where the multi-source data includes temperature mean, NO x Concentration variance, activity, trace elements, major chemical components, SO2 concentration, etc. The middle layer of the LSTM network structure is a two-layer stacked structure with 64 and 32 hidden units, respectively, and a dropout rate of 0.2. The output layer of the LSTM network structure can output three objectives: remaining life, denitrification efficiency, and poisoning risk level.

[0052] For example, the training process for the LSTM network structure includes:

[0053] Dataset: Historical data is divided into training set and test set with a ratio of 7:3;

[0054] Hyperparameters: learning rate 0.001, batch size 32, iterations 100;

[0055] Loss function: MAE is used for remaining life, MSE is used for denitrification efficiency, and cross entropy is used for poisoning risk.

[0056] For example, the incremental learning strategy is an adaptive update mechanism that includes:

[0057] Sliding window: retain the data of the last 30 days as the incremental training set;

[0058] Online learning: Model fine-tuning is automatically triggered at midnight every night, with an update cycle of ≤24 hours;

[0059] According to some embodiments of the present invention, referring to Figure 1 and Figure 2 ,The intelligent query module includes: natural language processing unit, multi-condition filter and visualization engine.

[0060] The natural language processing unit is used to parse user-entered query statements. For example, the intelligent query module includes a user input area. The natural language processing unit uses the BERT model to extract entities from the user input content (such as "Catalyst A" and "Activity <30m / h") and map them to database fields (Model = 'A', Activity <30m / h).

[0061] The multi-condition filter has a menu that includes query options, including the number of reactor layers, temperature range, operating time, and NOx removal efficiency threshold. The number of reactor layers has a submenu that is used to select the catalyst model. For example, the number of reactor layers can be selected as the first layer, the second layer, or the third layer; the temperature range can be selected as less than 400°C, the operating time can be selected as 0-20,000 hours, and NO x The removal efficiency threshold can be set to ≥80%. Catalyst models can be selected from plate, honeycomb, or corrugated plate. For example, the front-end interface provides a combination filter and supports SQL statement generation, such as: temperature between 300 and 400°C, catalyst operating time > 10,000 hours, etc.

[0062] The visualization engine is used to generate dynamic charts, including a line chart that overlays historical efficiency trends with forecast curves, a heat map showing the three-dimensional relationship between temperature, SO2 concentration, and efficiency, and a maintenance recommendation pop-up window. The triggering conditions for the maintenance recommendation pop-up window are predicted life < 15 days or efficiency degradation rate > 5% / month.

[0063] For example, the visualization engine can generate a trend chart and display it in the visualization display area: the ECharts library plots the denitrification efficiency historical curve and the prediction interval (95% confidence level).

[0064] For example, the visualization engine can generate a heat map and display it in the visualization display area: D3.js generates a three-dimensional matrix of temperature-SO2 concentration-efficiency, where red indicates high risk (efficiency <80%).

[0065] For example, the visualization engine can generate maintenance recommendations and display them in the visualization display area: the rule engine triggers prompts, such as "activity < 30m / h for 48h, it is recommended to start the regeneration program."

[0066] For example, the intelligent query module includes functional buttons, including data export and alarm settings. The data export function can export data to formats such as Word, PDF, and Excel. The alarm setting function allows you to customize the prediction efficiency threshold, triggering an alert when the prediction efficiency falls below the threshold. For example, the prediction efficiency threshold can be 85%.

[0067] According to some embodiments of the present invention, a dynamic data fusion-based intelligent query system for denitrification catalyst performance indicators includes a blockchain evidence storage unit, which includes data on-chain and a smart contract. The data on-chain is used to write at least part of the prediction results and maintenance operation records to a private blockchain node. The smart contract is used to generate an unalterable work order and assign it to the relevant personnel when preset conditions are triggered. By including the blockchain evidence storage unit in the dynamic data fusion-based intelligent query system for denitrification catalyst performance indicators, data encryption is achieved, improving data security and facilitating accountability for data modifications.

[0068] The catalyst performance index intelligent query method of the denitration catalyst performance index intelligent query system according to the second embodiment of the present invention includes:

[0069] S1. Multi-source data acquisition: The data acquisition module obtains the real-time operating parameters of the catalyst, the stored data in the laboratory database, and the environmental parameters;

[0070] S2. Dynamic data fusion: The dynamic data fusion module cleans, assigns weights, and filters noise from the data acquired by the data acquisition module to generate a fused data set.

[0071] S3. Performance prediction: Input the fused data set into the performance prediction model to calculate the catalyst life, efficiency and risk level;

[0072] S4. Result feedback: Return query results through the interactive interface.

[0073] For example, in a 1000MW coal-fired power plant SCR system, the catalyst performance index intelligent query method of the denitration catalyst performance index intelligent query system includes:

[0074] Data collection: Acquisition of data from data sensors, laboratory databases and environmental monitoring units;

[0075] Input query: "Expected life of catalyst A at loading > 90%";

[0076] The system returns: Remaining life is 5000h. The efficiency prediction curve shows a 1.6% decrease in the next 2500h. It is recommended to "re-check active ingredients in 4000h".

[0077] According to the catalyst performance index intelligent query method of the embodiment of the present invention, by including dynamic data fusion and performance prediction, the system can meet the needs of real-time and precise management, enhance the system's ability to cope with complex working conditions, and significantly improve catalyst management efficiency.

[0078] According to some embodiments of the present invention, referring to Figure 1-Figure 3 , the dynamic data fusion module cleans, assigns weights and filters noise on the data acquired by the data acquisition module to generate a fused data set, including:

[0079] Data cleaning: Eliminate values ​​that exceed the physical range in the data obtained by the data acquisition module;

[0080] Feature extraction: Select features that are strongly correlated with catalyst performance, including temperature mean, SO2 concentration peak, and active component decay rate;

[0081] Weight optimization: Entropy weight method weights are recalculated every 24 hours to adapt to changes in data source quality.

[0082] By cleaning, weighting and noise filtering the data obtained by the data acquisition module to generate a fused data set, the data in the fused data set can be made more accurate, more suitable for practical applications, and meet the needs of real-time and precise management.

[0083] According to some embodiments of the present invention, referring to Figure 1-Figure 3 The catalyst performance index intelligent query method of the denitration catalyst performance index intelligent query system also includes storing the output results of the performance prediction model through the blockchain. Storing the output results of the performance prediction model through the blockchain includes:

[0084] Data hashing: Generate a unique hash value for the prediction result and write it into the blockchain distributed ledger;

[0085] Evidence verification: Verify whether historical prediction records have been tampered with through the query interface.

[0086] By storing the output results of the performance prediction model through blockchain, the output results of the prediction model can be accurately recorded, improving the security of the output data and making it easier to hold relevant personnel accountable after making changes to the data.

[0087] Refer to the following Figure 3 The following describes a catalyst performance index intelligent query method of a denitration catalyst performance index intelligent query system according to some specific embodiments of the present invention:

[0088] S10: Data input: sensor data, laboratory data, environmental data.

[0089] S11: Data cleaning: remove outliers and fill in missing data.

[0090] S12: Feature extraction: Extract key indicator features such as SO2.

[0091] S13: Entropy weight method weight allocation.

[0092] S14: Kalman filter denoising: Smoothing out the temporal noise of sensor data.

[0093] S15: Data fusion output: Generate a standardized and weighted comprehensive data set.

[0094] The final output is the prediction model.

[0095] In the description of the present invention, "first feature" or "second feature" may include one or more of the features.

[0096] In the description of the present invention, "plurality" means two or more.

[0097] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative uses of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0098] While 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 variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. An intelligent query system for denitrification catalyst performance indicators based on dynamic data fusion, characterized in that: include: Data acquisition module, used to obtain real-time operating parameters of the catalyst, stored data in the laboratory database, and environmental parameters; A dynamic data fusion module is electrically connected to the data acquisition module and is used to process the data acquired by the data acquisition module. The dynamic data fusion module includes a Kalman filter unit and an entropy weight method weight unit. The dynamic data fusion module performs weight assignment, noise filtering and standardization on the data acquired by the data acquisition module based on the entropy weight method and the Kalman filter algorithm to generate a fused data set; a performance prediction module having a performance prediction model and configured to receive the generated fusion data set generated by the dynamic data fusion module, wherein the performance prediction model is configured to predict and output life prediction value data, denitration efficiency attenuation trend, and poisoning risk level of the catalyst; An intelligent query module includes an interactive interface for querying catalyst performance indicators through natural language input and multi-condition screening, and displaying them in a visual interface.

2. The denitration catalyst performance index intelligent query system based on dynamic data fusion according to claim 1 is characterized in that: The data acquisition module is used to be electrically connected to the distributed control system to collect real-time operating parameters of the catalyst in the distributed control system, wherein the real-time operating parameters of the catalyst include temperature, pressure, NOx concentration, and ammonia escape rate; The data acquisition module is used to be electrically connected to the laboratory database to obtain the stored data of the laboratory database, wherein the stored data of the laboratory database includes specific surface area, chemical composition, trace elements and activity; The data acquisition module is used to be electrically connected to the environmental monitoring equipment to obtain the environmental parameters monitored by the environmental monitoring equipment. The environmental monitoring equipment includes a flue gas analyzer and a humidity sensor to monitor SO2 concentration, humidity and oxygen content. The environmental parameters include SO2 concentration, humidity and oxygen content.

3. The intelligent query system for denitration catalyst performance indicators based on dynamic data fusion according to claim 1 is characterized in that: An API interface is included, through which the denitration catalyst performance index intelligent query system is integrated with an external control system. The API interface includes a data input interface and a data output interface. The data input interface is used to receive real-time operating condition data from a distributed control system, and the data output interface is used to push a catalyst replacement warning signal to the manufacturing execution system. And / or, the dynamic data fusion module has the function of supporting offline-real-time data synchronization and includes offline data loading and real-time data stream processing, wherein the offline data loading includes batch importing historical laboratory test reports and extracting structured data through optical character recognition technology, and the real-time data stream processing includes using Apache Kafka to build a message queue to achieve low-latency transmission and buffering of sensor data, and the low-latency delay time is less than 1 second.

4. The denitration catalyst performance index intelligent query system based on dynamic data fusion according to claim 1 is characterized in that: The entropy weight method weight unit is used to calculate the weight coefficient according to the reliability, update frequency and correlation of the data source with the catalyst performance. The formula for calculating the weight coefficient is: Among them, V a is the entropy value of the a-th data, p ij is the proportion of data after standardization; The Kalman filter unit is used to construct a state space model based on the data collected by the data acquisition module, and eliminate time series noise through a prediction-correction cycle. The state equation and observation equation are: k =Ax k-1 +w k-1 , z k =Hx k +v k , where x k is the system state vector, Z k is the observed value, W k 、V k are process noise and observation noise.

5. The intelligent query system for denitration catalyst performance indicators based on dynamic data fusion according to claim 1 is characterized in that: The performance prediction model includes an LSTM network structure, which includes: an input layer, two LSTM hidden layers, and an output layer. The input layer is used to fuse data feature dimensions. The number of neurons in the two LSTM hidden layers is 64 / 32. The output data of the output layer includes the remaining life of the catalyst, the denitrification efficiency, and the degree of catalyst deactivation. The performance prediction model is configured to adopt an incremental learning strategy to dynamically adjust the weights in the entropy weight method weight unit.

6. The intelligent query system for denitration catalyst performance indicators based on dynamic data fusion according to claim 1 is characterized in that: The intelligent query module includes: A natural language processing unit, configured to parse a query statement input by a user; A multi-condition filter, wherein the multi-condition filter has a menu including a query scheme, the query scheme including the number of reactor layers, temperature range, operating time, and NOx removal efficiency threshold, and a submenu under the number of reactor layers, the submenu being used to select a catalyst model; A visualization engine is used to generate dynamic charts, including a line chart that overlays historical efficiency trends with forecast curves, a heat map of the three-dimensional relationship between temperature, SO2 concentration, and efficiency, and a maintenance suggestion pop-up window. The triggering conditions for the maintenance suggestion pop-up window are predicted life < 15 days or efficiency degradation rate > 5% / month.

7. The intelligent query system for denitration catalyst performance indicators based on dynamic data fusion according to claim 1 is characterized in that: It includes a blockchain evidence storage unit, which includes data on the chain and a smart contract. The data on the chain is used to write at least part of the prediction results and maintenance operation records into a private chain node. The smart contract is used to generate an unalterable work order and assign it to relevant personnel when the preset conditions are triggered.

8. A catalyst performance index intelligent query method based on the denitration catalyst performance index intelligent query system according to any one of claims 1 to 7, characterized in that: include: S1. Multi-source data acquisition: The data acquisition module acquires the real-time operating parameters of the catalyst, the stored data in the laboratory database, and the environmental parameters; S2. Dynamic data fusion: The dynamic data fusion module cleans, assigns weights, and filters noise on the data acquired by the data acquisition module to generate the fused data set; S3. Performance prediction: inputting the fused data set into the performance prediction model to calculate the catalyst life, efficiency and risk level; S4. Result feedback: returning the query result through the interactive interface.

9. The catalyst performance index intelligent query method according to claim 8, characterized in that: The data acquired by the data acquisition module is cleaned, weighted, and noise filtered by the dynamic data fusion module to generate the fused data set, including: Data cleaning: eliminating values ​​exceeding the physical range in the data acquired by the data acquisition module; Feature extraction: Select features that are strongly correlated with catalyst performance, including temperature mean, SO2 concentration peak, and active component decay rate; Weight optimization: Entropy weight method weights are recalculated every 24 hours to adapt to changes in data source quality.

10. The catalyst performance index intelligent query method according to claim 9, characterized in that: The method further includes storing the output result of the performance prediction model through a blockchain. Storing the output result of the performance prediction model through a blockchain includes: Data hashing: Generate a unique hash value for the prediction result and write it into the blockchain distributed ledger; Evidence verification: Verify whether historical prediction records have been tampered with through the query interface.

Images