An intelligent recommendation method for chemical products based on data analysis

Abstract

The present application provides a method for intelligent recommendation of chemical products based on data analysis, which involves data recommendation, including: acquiring historical storage data and current production data of a chemical plant, establishing a chemical product database according to the historical storage data of the chemical plant, performing feature extraction on the current production data to obtain feature extraction data, and at the same time, performing data analysis on the data in the chemical product database to construct a data knowledge base of chemical products, respectively using a chemical output prediction model and a chemical quality prediction model to predict the output prediction results and quality prediction results of chemical products in the current state, and then inputting the feature extraction data, the quality prediction results and the output prediction results of chemical products in the current state into a recommendation system, and obtaining recommended production parameters and operation strategies of chemical products according to the data knowledge base of chemical products, and displaying them on a chemical production recommendation interface, thereby ensuring the quality stability and consistency of chemical products.

Description

Technical Field

[0001] The present application relates to the technical field of data recommendation, and in particular to a method for intelligently recommending chemical products based on data analysis. Background Art

[0002] Chemical plants face the dual problems of output and quality during the production process. The output problem involves how to maximize the number of products produced, while the quality problem focuses on how to ensure that the products meet the expected standards and requirements. Among them, the output problem may be caused by a variety of factors, such as the quality and supply stability of chemical raw materials, the adjustment of process conditions, the selection of catalysts, etc. If not properly controlled or adjusted, it may lead to a decrease in output, resulting in low production efficiency and losses. The quality problem is also crucial for chemical plants. The quality of the product is directly related to the market competitiveness and customer satisfaction.

[0003] Therefore, chemical plants need to find a way to ensure the quality stability and consistency of their products, thereby improving production efficiency, reducing costs, achieving stable and reliable operation of chemical plants, and gaining an advantage in the highly competitive market. Summary of the invention

[0004] The present invention provides a chemical product intelligent recommendation method based on data analysis, which is used to ensure the quality stability and consistency of chemical products.

[0005] The present invention provides a method for intelligently recommending chemical products based on data analysis. The method comprises:

[0006] Obtain the historical storage data and current production data of the chemical plant, and establish a chemical product database based on the historical storage data of the chemical plant;

[0007] Extract features from current production data to obtain feature extraction data. At the same time, analyze the data in the chemical product database to build a data knowledge base for chemical products.

[0008] The feature extraction data is predicted using the chemical production prediction model to obtain the production prediction results of chemical products in the current state, and the feature extraction data is predicted using the chemical quality prediction model to obtain the quality prediction results of chemical products in the current state;

[0009] The feature extraction data, the quality prediction results and the output prediction results of the chemical products in the current state are input into the recommendation system. According to the data knowledge base of the chemical products, the recommended production parameters and operation strategies of the chemical products are obtained and displayed on the chemical production recommendation interface.

[0010] In one possible design, the chemical product database includes the raw material composition, process parameters, catalyst usage and product indicators of the chemical products;

[0011] Raw material composition includes the raw material composition, proportion and supplier information used in each production batch of chemical products;

[0012] Process parameters include temperature, pressure and reaction time during the production of each production batch of chemical products;

[0013] Catalyst usage includes the type and amount of catalyst used in the production process of each production batch of chemical products;

[0014] Product indicators include the output index value and quality index value of each production batch of chemical products.

[0015] Optionally, feature extraction is performed on the current production data to obtain feature extraction data, including:

[0016] Extract the raw material composition and proportion of chemical products, temperature, pressure, reaction time and catalyst usage during the production process from current production data;

[0017] Standardize the raw material composition, process parameters and catalyst usage of chemical products to obtain standardized chemical product data characteristics;

[0018] The principal component analysis is performed on the standardized chemical product data features, and the standardized chemical product data features are mapped to a unified feature space to obtain feature extraction data.

[0019] Optionally, data analysis is performed on the data in the chemical product database to construct a data knowledge base of chemical products, including:

[0020] Analyze the production impact of the raw material composition and catalyst usage of chemical products on product indicators based on the raw material composition, catalyst usage and product quality indicators of chemical products;

[0021] According to the process parameters in the chemical product database, the variation range of process parameters of chemical products and the correlation between process parameters and product indicators are analyzed.

[0022] In a possible design, the chemical production prediction model is used to predict the feature extraction data to obtain the production prediction results of chemical products under the current state, including:

[0023] Input the feature extraction data into the chemical production prediction model;

[0024] The chemical production prediction model calculates the output prediction value of chemical products in the current state based on the process parameters in the feature extraction data;

[0025] The output prediction value of the chemical product under the current state is normalized to obtain the output prediction result of the chemical product under the current state.

[0026] Optionally, the chemical production prediction model is a feed-forward neural network model;

[0027] The feedforward neural network model is pre-trained by the process parameters and output index values ​​in the chemical product database.

[0028] In a possible design, the chemical quality prediction model is used to predict the feature extraction data to obtain the quality prediction results of the chemical products in the current state, including:

[0029] Input the feature extraction data into the chemical quality prediction model;

[0030] The chemical quality prediction model calculates the quality prediction value of chemical products in the current state based on the raw material composition and catalyst usage in the feature extraction data;

[0031] The quality prediction value of the chemical product in the current state is normalized to obtain the quality prediction result of the chemical product in the current state.

[0032] Optionally, the chemical quality prediction model is an XGBoost model.

[0033] The XGBoost model is pre-trained based on the raw material composition, catalyst usage, and quality index values ​​in the chemical product database.

[0034] In a possible design, feature extraction data, quality prediction results and output prediction results of chemical products in the current state are input into the recommendation system, and the recommended production parameters and operation strategies of chemical products are obtained according to the data knowledge base of chemical products, including:

[0035] When the difference between the quality prediction result of the chemical product in the current state and the preset quality exceeds a first preset difference, a corresponding recommended raw material component value and a recommended catalyst dosage are generated according to the data knowledge base of the chemical product;

[0036] When the difference between the output prediction result of the chemical product in the current state and the preset output exceeds a second preset difference, corresponding recommended process parameters are generated according to the data knowledge base of the chemical product.

[0037] In a possible design, the method further includes:

[0038] Conduct safety monitoring of process parameters and catalyst usage of chemical products;

[0039] When the temperature, pressure and reaction time in the process parameters of chemical products exceed the corresponding preset range, the first safety warning will be issued on the chemical production recommendation interface;

[0040] When the catalyst usage in the catalyst usage of a chemical product exceeds the preset usage threshold, a second safety warning is issued on the chemical production recommendation interface.

[0041] The present invention provides an intelligent recommendation method for chemical products based on data analysis, the method comprising: acquiring historical storage data and current production data of a chemical plant, establishing a chemical product database according to the historical storage data of the chemical plant, performing feature extraction on the current production data to obtain feature extraction data, and at the same time, performing data analysis on the data in the chemical product database to construct a data knowledge base for chemical products, using a chemical output prediction model to predict the feature extraction data to obtain an output prediction result of the chemical product in the current state, using a chemical quality prediction model to predict the feature extraction data to obtain a quality prediction result of the chemical product in the current state, inputting the feature extraction data, the quality prediction result of the chemical product in the current state, and the output prediction result into a recommendation system, obtaining recommended production parameters and operation strategies for the chemical products according to the data knowledge base for the chemical products, and displaying them on a chemical production recommendation interface. That is to say, by obtaining the historical storage data of the chemical plant and establishing a chemical product database, we can make full use of past production records and data to form a valuable knowledge base, which provides a basis for subsequent data analysis and prediction. By analyzing the data in the chemical product database, we can reveal the relationship between different characteristics and output and quality, and then build a data knowledge base for chemical products to provide a basis for subsequent predictions and recommendations. With the help of chemical output prediction models and quality prediction models, the feature extraction data is predicted, and then combined with the data knowledge base activities to recommend production parameters and operating strategies, thereby improving the decision-making ability and work efficiency of operators and ensuring the quality stability and consistency of chemical products.

[0042] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The above and other features, advantages and aspects of the embodiments of the present invention will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements.

[0044] Figure 1 This is a flow chart of a chemical product intelligent recommendation method based on data analysis according to an embodiment of the present invention;

[0045] Figure 2 A flowchart of obtaining feature extraction data according to an embodiment of the present invention;

[0046] Figure 3A schematic diagram of a structure for generating a yield prediction result and a quality prediction result according to an embodiment of the present invention;

[0047] Figure 4 is a schematic diagram of the structure of a recommendation system according to an embodiment of the present invention;

[0048] Figure 5 It is a structural schematic diagram of the chemical production recommendation interface of an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0050] It should be noted that the description of the embodiments of the present invention is only for more clearly illustrating the technical solutions of the embodiments of the present invention, and does not constitute a limitation on the technical solutions provided by the embodiments of the present invention.

[0051] The embodiment of the present invention provides a chemical product intelligent recommendation method based on data analysis. Figure 1 Flow chart of the quality assessment model training method according to an embodiment of the present invention. Figure 1 , the method comprising:

[0052] S101. Acquire historical storage data and current production data of a chemical plant, and establish a chemical product database based on the historical storage data of the chemical plant.

[0053] In this embodiment, by acquiring the historical storage data of the chemical plant and establishing a chemical product database, the historical data can be fully utilized to provide data support, provide a basis for subsequent data analysis and prediction of current production data, and better realize intelligent recommendations in the chemical product production process.

[0054] Optionally, the chemical product database includes raw material composition, process parameters, catalyst usage and product indicators of the chemical products;

[0055] Raw material composition includes the raw material composition, proportion and supplier information used in each production batch of chemical products;

[0056] Process parameters include temperature, pressure and reaction time during the production of each production batch of chemical products;

[0057] Catalyst usage includes the type and amount of catalyst used in the production process of each production batch of chemical products;

[0058] Product indicators include the output index value and quality index value of each production batch of chemical products.

[0059] The chemical product database may include the raw material composition, process parameters, catalyst usage and product indicators of multiple different chemical products in each production batch. For example, if the products produced by a chemical plant include coating chemical products, then the chemical product database may include the raw material composition, process parameters, catalyst usage and product indicators of coating chemical products, wherein each coating chemical product has a unique product number and name, and the chemical product database may also include information descriptions of the coating product, such as hazard assessment, safe operation guidelines and first aid measures.

[0060] In this embodiment, the raw material composition, process parameters, catalyst usage and product indicators of chemical products are stored in the chemical product database, providing comprehensive chemical product information, including historical production data, which can trace the production conditions of different batches and facilitate the analysis of chemical product production data.

[0061] S102. Perform feature extraction on current production data to obtain feature extraction data. At the same time, perform data analysis on the data in the chemical product database to build a data knowledge base for chemical products.

[0062] In this embodiment, by performing feature extraction on the current production data, representative and key features can be extracted from the complex original data, data analysis can be performed on the data in the chemical product database, and a data knowledge base of chemical products can be constructed, which helps to form a comprehensive and valuable knowledge base and provide a basis for intelligent recommendations in the subsequent chemical product production process.

[0063] Optionally, feature extraction is performed on the current production data to obtain feature extraction data, including:

[0064] Extract the raw material composition and proportion of chemical products, temperature, pressure, reaction time and catalyst usage during the production process from current production data;

[0065] Standardize the raw material composition, process parameters and catalyst usage of chemical products to obtain standardized chemical product data characteristics;

[0066] The principal component analysis is performed on the standardized chemical product data features, and the standardized chemical product data features are mapped to a unified feature space to obtain feature extraction data.

[0067] Figure 2FIG. 1 is a flow chart of obtaining feature extraction data according to an embodiment of the present invention, as shown in FIG. Figure 2 shown.

[0068] Among them, each variable in the standardized chemical product data features has zero mean and unit variance to eliminate the dimensional differences between different variables. The principal component analysis decomposes the standardized chemical product data features to obtain eigenvalues ​​and corresponding eigenvectors. According to the size of the eigenvalues, the eigenvectors corresponding to the first few eigenvalues ​​are selected as the principal components, and the standardized chemical product data features are mapped to a unified feature space to obtain feature extraction data. For example, the current production data of a chemical product produced by a chemical plant includes raw material composition, process parameters and catalyst usage. The raw material composition includes raw material A, raw material B and raw material C, with the proportions of raw material A accounting for 50%, raw material B accounting for 30%, and raw material C accounting for 20%. The temperature in the production process is 200°C, the pressure is 5 bar, the reaction time is 2 hours, and the amount of catalyst X is 10g. Then, the chemical product data features after standardization of the current production data are [[0.5, 0.3, 0.2], [200, 5, 2],

[10] ], and the feature extraction data after principal component analysis of the standardized chemical product data features are [[0.1, 0.2], [0.3, 0.2], [0.5]], and the feature extraction data is obtained, which effectively realizes the data dimensionality reduction and mapping of the standardized chemical product data features.

[0069] In this embodiment, the raw material composition and proportion of chemical products, the temperature, pressure, reaction time and catalyst usage in the production process are extracted from the current production data, standardized, and the standardized chemical product data features are subjected to principal component analysis and mapped to a unified feature space. This can reduce the dimensionality of the chemical product data and extract the most representative features, thereby reducing redundant information and providing a more concise and efficient data basis for subsequent data prediction analysis and recommendation systems.

[0070] Optionally, data analysis is performed on the data in the chemical product database to construct a data knowledge base of chemical products, including:

[0071] Analyze the production impact of the raw material composition and catalyst usage of chemical products on product indicators based on the raw material composition, catalyst usage and product quality indicators of chemical products;

[0072] According to the process parameters in the chemical product database, the variation range of process parameters of chemical products and the correlation between process parameters and product indicators are analyzed.

[0073] Among them, the raw material composition of chemical products, the use of catalysts and the product quality indicators can be analyzed respectively to analyze the production impact of the raw material composition and catalyst use of chemical products on product indicators. For example, when the chemical product is ethylene, the raw material composition and product quality indicators of ethylene in the chemical product database can be analyzed to determine the impact of different raw material components on the quality of ethylene products, including but not limited to the impact of ethane content, propane content and impurity content in ethylene on product indicators. Assuming that the ideal purity of ethylene is above 99.5%, after data analysis, it is found that for every 1% increase in ethane content in ethylene, the purity of ethylene decreases by about 0.2%. That is to say, when the ethane content in ethylene is 2%, the purity of ethylene may be reduced to 99.1%. Then, the influence relationship of ethane content in ethylene on product indicators can be established and recorded in the data knowledge base for next use. Similarly, the influence of catalysts on product indicators can be determined by analyzing the use of catalysts used in the production process of ethylene in the chemical product database, thereby helping enterprises to select the best catalyst and catalyst dosage to improve the output and quality of ethylene.

[0074] Among them, the process parameter variation range of chemical products and the correlation between process parameters and product indicators are analyzed according to the process parameters in the chemical product database, including but not limited to temperature, pressure and reaction time. For example, the production temperature variation range of chemical products is usually 150°C to 300°C. The increase in temperature can increase the reaction rate, thereby increasing the output of chemical products. The pressure range is usually 1 to 10 MPa, and increasing the reaction pressure can increase the concentration of reactants, promote the reaction, and also increase the output of chemical products. The reaction time range varies greatly according to different chemical products. The specific process parameter variation range and the correlation between process parameters and product indicators vary due to different chemical processes and reactions. Some chemical products can be more fully converted under longer reaction times, which can also increase the output of chemical products. Therefore, by analyzing the variation range of process parameters and the correlation with product indicators, it is helpful to increase the output of chemical products on the production line of chemical plants.

[0075] In this embodiment, by analyzing the data in the chemical product database and constructing a data knowledge base of chemical products, chemical plants can be helped to achieve precise product quality control, provide optimal production parameter selection, and help improve the production efficiency of chemical plants, the quality of chemical products and resource utilization, thereby enhancing the production competitiveness of chemical plants.

[0076] S103. Use a chemical production prediction model to predict the feature extraction data to obtain a production prediction result of the chemical product in the current state, and use a chemical quality prediction model to predict the feature extraction data to obtain a quality prediction result of the chemical product in the current state.

[0077] Figure 3 FIG. 1 is a schematic diagram of a structure for generating a yield prediction result and a quality prediction result according to an embodiment of the present invention. Figure 3 shown.

[0078] In this embodiment, the chemical production prediction model and the chemical quality prediction model are used to perform production prediction and quality prediction on chemical products respectively, so that each model can learn a specific set of knowledge and goals more focusedly, thereby achieving more accurate production prediction and quality prediction.

[0079] Optionally, the feature extraction data is predicted using a chemical production prediction model to obtain a production prediction result of chemical products in the current state, including:

[0080] Input the feature extraction data into the chemical production prediction model;

[0081] The chemical production prediction model calculates the output prediction value of chemical products in the current state based on the process parameters in the feature extraction data;

[0082] The output prediction value of the chemical product under the current state is normalized to obtain the output prediction result of the chemical product under the current state.

[0083] In this embodiment, the feature extraction data is predicted using a chemical production prediction model, which can provide efficient and accurate production prediction results, thereby helping chemical plants optimize production management and increase the production output of chemical products.

[0084] Optionally, the chemical production prediction model is a feed-forward neural network model;

[0085] The feedforward neural network model is pre-trained by the process parameters and output index values ​​in the chemical product database.

[0086] Among them, the feedforward neural network is a powerful nonlinear model that can handle complex nonlinear relationships. In chemical production, since the output is often affected by multiple factors, there may be complex nonlinear relationships between these factors, and the feedforward neural network can accept multi-dimensional input, that is, multiple factors can be used as input features for modeling. For example, there are 3 process parameters of chemical products, that is, there are 3 input features of the feedforward neural network, and the neuron to be output is a yield index value, then a feedforward neural network model with 2 hidden layers can be set, each hidden layer contains 10 neurons, where the input layer is 3 neurons (corresponding to 3 process parameters), hidden layer 1 is 10 neurons, hidden layer 2 is 10 neurons, and the output layer is 1 neuron (corresponding to the yield index value), thereby realizing the output prediction of chemical products. Therefore, the feedforward neural network takes the process parameters in the chemical product database as input features, comprehensively considers multiple factors, and thus improves the accuracy of yield prediction. By training historical yield data, the feedforward neural network can learn the degree and weight of the influence of different factors on yield, so as to accurately predict future yields.

[0087] In this embodiment, the feedforward neural network model is used as a tool for predicting chemical product output. It has the advantages of adaptability, nonlinear modeling capability, multi-feature processing, scalability and real-time performance. The feedforward neural network model is pre-trained with process parameters and output index values ​​in the chemical product database, which effectively improves the accuracy of the output prediction results of the feedforward neural network model.

[0088] Optionally, the chemical quality prediction model is used to predict the feature extraction data to obtain the quality prediction results of the chemical products in the current state, including:

[0089] Input the feature extraction data into the chemical quality prediction model;

[0090] The chemical quality prediction model calculates the quality prediction value of chemical products in the current state based on the raw material composition and catalyst usage in the feature extraction data;

[0091] The quality prediction value of the chemical product in the current state is normalized to obtain the quality prediction result of the chemical product in the current state.

[0092] In this embodiment, the feature extraction data is predicted using a chemical quality prediction model, which can provide efficient and accurate quality prediction results, thereby helping chemical plants optimize production management and improve the production quality of chemical products.

[0093] Optionally, the chemical quality prediction model is an XGBoost model.

[0094] The XGBoost model is pre-trained based on the raw material composition, catalyst usage, and quality index values ​​in the chemical product database.

[0095] For example, in the chemical product database, there are 10 raw material components in a chemical product and 3 types of catalyst usage in the production process. These data are used as input features, and the quality index value in the chemical product database is used as the target variable. The data input of the raw material component part is [0.2, 0.3, 0.5, 0.1, 0.4, 0.6, 0.7, 0.8, 0.9, 0.5], and the input of the catalyst usage part is [1, 0, 1]. The quality index value in the chemical product database is 8. The quality prediction result output by the XGBoost model is calculated with the quality index value, and the training parameters of the XGBoost model are corrected so that the model can perform better predictions next time. Therefore, the XGBoost model is continuously trained using the data in the chemical product database to optimize the XGBoost model's quality prediction of chemical products, and then an XGBoost model with high prediction accuracy can be obtained.

[0096] Among them, the XGBoost (eXtreme Gradient Boosting) model is a gradient boosting decision tree model. It builds a powerful prediction model by integrating multiple weak classifiers (decision trees). It can handle nonlinear relationships and interactions between features and has the ability to handle high-dimensional sparse data. In chemical quality prediction, the input features are usually high-dimensional and sparse, that is, the feature dimensions are high and there are a large number of zero values ​​between features. The feedforward neural network model may face difficulties in processing high-dimensional sparse data. Therefore, using the XGBoost model for quality prediction of chemical products will achieve better results.

[0097] In this embodiment, the XGBoost model is used as a tool for chemical product quality prediction, and has the advantages of high performance, accuracy, feature importance evaluation, robustness and interpretability. The XGBoost model is pre-trained by the raw material composition, catalyst usage and quality index values ​​in the chemical product database, which effectively improves the prediction accuracy of the XGBoost model and realizes real-time optimization of the XGBoost model.

[0098] S104. Input the feature extraction data, the quality prediction results and the output prediction results of the chemical products in the current state into the recommendation system, obtain the recommended production parameters and operation strategies of the chemical products according to the data knowledge base of the chemical products, and display them on the chemical production recommendation interface.

[0099] In this embodiment, feature extraction data, quality prediction results and output prediction results of chemical products are input into the recommendation system, and combined with the data knowledge base of chemical products to realize the recommendation of production parameters and operating strategies in the production process of chemical products, which accelerates the decision-making process, optimizes resource utilization and cost control, and helps to enhance the production efficiency, product quality and competitiveness of chemical plants.

[0100] Optionally, the feature extraction data, the quality prediction results and the output prediction results of the chemical products in the current state are input into the recommendation system, and the recommended production parameters and operation strategies of the chemical products are obtained according to the data knowledge base of the chemical products, including:

[0101] When the difference between the quality prediction result of the chemical product in the current state and the preset quality exceeds a first preset difference, a corresponding recommended raw material component value and a recommended catalyst dosage are generated according to the data knowledge base of the chemical product;

[0102] When the difference between the output prediction result of the chemical product in the current state and the preset output exceeds a second preset difference, corresponding recommended process parameters are generated according to the data knowledge base of the chemical product.

[0103] Among them, when the difference between the quality prediction result and the preset quality of the chemical product in the current state exceeds the first preset difference, the corresponding recommended raw material component value and the recommended catalyst dosage can be generated according to the data knowledge base of the chemical product. For example, the chemical product is a coating, in which the raw material components include pigment A, pigment B and diluent C. The preset quality of the chemical product is 100 units, and the actual quality prediction result is 120 units, which exceeds the first preset difference (assuming that the first preset difference is 10 units). The example records between the raw material components and the quality units in the data knowledge base are as follows:

[0104] Pigment A Pigment B Diluent C Unit of mass 20 30 50 105 25 35 45 95 18 32 50 40 ... ... ... ...

[0105] Based on the records and analysis of multiple pieces of information in the data knowledge base, the recommendation system believes that increasing pigment A and reducing diluent C can make the mass unit of the chemical product close to the preset mass. Furthermore, based on the difference between the current mass prediction result of the chemical product and the preset mass unit, the recommended raw material component values ​​can be generated as follows: the recommended pigment A is 25 units, the recommended pigment B is 30 units, and the recommended diluent C is 45 units. Then, the chemical plant can adjust the raw material ratio of the chemical product according to this recommendation. Similarly, for example, the chemical product is a coating, and 9g of catalyst is used in the production process. The preset mass of the chemical product is 100 units, and the actual mass prediction result is 120 units, which exceeds the first preset difference (assuming that the first preset difference is 10 units). The example records between the catalyst dosage and the mass unit in the data knowledge base are as follows:

[0106] Catalyst dosage / g Unit of mass 9 120 7 100 6 90 ... ...

[0107] Then, based on the analysis, the recommendation system believes that reducing the amount of catalyst can bring the mass closer to the preset value. Therefore, the recommendation system may generate a recommended catalyst amount of 7 grams.

[0108] When the difference between the output prediction result and the preset output of the chemical product in the current state exceeds the second preset difference, the corresponding recommended process parameters can be generated according to the data knowledge base of the chemical product. For example, the chemical product is a polymer, the preset output is 1000 kg, and the output prediction result is 1200 kg, which exceeds the second preset difference (assuming that the second preset difference is 100 kg). The example records between the process parameters in the data knowledge base and the output of the chemical product are as follows:

[0109] Temperature / ℃ Pressure / MPa Response time / hour Output / kg 80 2 2 900 85 2.5 2.5 1050 75 1.8 1.5 800 ... ... ... ...

[0110] Based on the analysis, the recommendation system believes that increasing the temperature and extending the reaction time can increase the yield. Therefore, the recommendation system may generate the following recommended process parameters: temperature of 90°C, pressure of 2.5MPa, and reaction time of 3 hours.

[0111] Figure 4 Schematic diagram of the structure of the recommendation system of the embodiment of the present invention. Figure 4 shown.

[0112] In this embodiment, the corresponding recommended raw material composition values, recommended catalyst dosages and recommended process parameters are generated based on the data knowledge base of chemical products, which can help chemical plants to achieve precise control of the quality and output of chemical products, thereby improving production efficiency, reducing costs and increasing profits.

[0113] Optionally, the method further comprises,

[0114] Conduct safety monitoring of process parameters and catalyst usage of chemical products;

[0115] When the temperature, pressure and reaction time in the process parameters of chemical products exceed the corresponding preset range, the first safety warning will be issued on the chemical production recommendation interface;

[0116] When the catalyst usage in the catalyst usage of a chemical product exceeds the preset usage threshold, a second safety warning is issued on the chemical production recommendation interface.

[0117] Figure 5 Schematic diagram of the structure of the chemical production recommendation interface of an embodiment of the present invention, such as Figure 5 shown.

[0118] The recommended interface for chemical production includes "temperature", "pressure", "reaction time", "catalyst usage", "first safety warning" and "second safety warning". Among them, "temperature", "pressure", "reaction time" and "catalyst usage" are used to display the corresponding parameter values. The "first safety warning" is used to issue a timely reminder when the temperature, pressure and reaction time in the process parameters of the chemical product exceed the corresponding preset range. The "second safety warning" is used to issue a timely reminder when the catalyst usage in the catalyst usage of the chemical product exceeds the preset usage threshold.

[0119] The preset range and preset dosage threshold can be set as needed. For example, when the temperature of the process parameters of the chemical product exceeds 300°C, the pressure exceeds 10MPa, and the reaction time exceeds 2 hours, a reminder can be given in the "first safety warning". When the actual catalyst dosage exceeds 100 grams, a reminder can be given in the "second safety warning".

[0120] In this embodiment, the process parameters of chemical products and the use of catalysts are safely monitored, and corresponding safety warnings are issued on the chemical production recommendation interface, which can ensure the safety of the chemical product production process and avoid accidents in chemical plants.

[0121] Based on the intelligent recommendation method for chemical products based on data analysis provided in this embodiment, historical storage data and current production data of the chemical plant are obtained, a chemical product database is established according to the historical storage data of the chemical plant, feature extraction is performed on the current production data to obtain feature extraction data, and at the same time, data analysis is performed on the data in the chemical product database to build a data knowledge base for chemical products, a chemical output prediction model is used to predict the feature extraction data to obtain the output prediction result of the chemical product in the current state, a chemical quality prediction model is used to predict the feature extraction data to obtain the quality prediction result of the chemical product in the current state, the feature extraction data, the quality prediction result of the chemical product in the current state, and the output prediction result are input into the recommendation system, and the recommended production parameters and operation strategies of the chemical products are obtained according to the data knowledge base of the chemical products, and are displayed on the chemical production recommendation interface. That is to say, by obtaining the historical storage data of the chemical plant and establishing a chemical product database, we can make full use of past production records and data to form a valuable knowledge base, which provides a basis for subsequent data analysis and prediction. By analyzing the data in the chemical product database, we can reveal the relationship between different characteristics and output and quality, and then build a data knowledge base for chemical products to provide a basis for subsequent predictions and recommendations. With the help of chemical output prediction models and quality prediction models, the feature extraction data is predicted, and then combined with the data knowledge base activities to recommend production parameters and operating strategies, thereby improving the decision-making ability and work efficiency of operators and ensuring the quality stability and consistency of chemical products.

[0122] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a different order from the order marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0123] The units or modules involved in the embodiments of the present invention may be implemented by software or hardware. The units or modules described may also be arranged in a processor, and the names of these modules do not constitute limitations on the modules themselves in some cases.

[0124] The above description is only a preferred embodiment of the present invention and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present invention is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present invention (but not limited to) to form a technical solution.

Claims

1. A chemical product intelligent recommendation method based on data analysis, characterized in that: The method comprises: Acquire historical storage data and current production data of a chemical plant, and establish a chemical product database based on the historical storage data of the chemical plant; Performing feature extraction on the current production data to obtain feature extraction data, and at the same time, performing data analysis on the data in the chemical product database to construct a data knowledge base of the chemical products; Using a chemical production prediction model to predict the feature extraction data, to obtain a production prediction result of the chemical product in the current state; using a chemical quality prediction model to predict the feature extraction data, to obtain a quality prediction result of the chemical product in the current state; Input the feature extraction data, the quality prediction results and the output prediction results of the chemical products in the current state into the recommendation system, obtain the recommended production parameters and operation strategies of the chemical products according to the data knowledge base of the chemical products, and display them on the chemical production recommendation interface; The chemical product database includes the raw material composition, process parameters, catalyst usage and product indicators of the chemical products; The raw material composition includes the raw material composition, proportion and supplier information used in each production batch of the chemical product; The process parameters include temperature, pressure and reaction time during the production of each production batch of the chemical product; The catalyst usage includes the type and amount of catalyst used in the production process of each production batch of the chemical product; The product index includes the output index value and quality index value of each production batch of the chemical product; The chemical product database also includes: product information description, which includes: risk assessment, safe operation guidelines and first aid measures; The step of extracting features from the current production data to obtain feature extraction data includes: Extracting the raw material composition and proportion of the chemical product, the temperature, pressure, reaction time and catalyst usage during the production process from the current production data; Standardizing the raw material composition, process parameters and catalyst usage of the chemical product to obtain standardized chemical product data characteristics; Performing principal component analysis on the standardized chemical product data features, mapping the standardized chemical product data features to a unified feature space, and obtaining the feature extraction data; The data analysis of the data in the chemical product database to construct the data knowledge base of the chemical products includes: According to the raw material composition of the chemical product, the use of the catalyst and the product index, respectively analyze the production impact of the raw material composition of the chemical product and the use of the catalyst on the product index; Analyzing the variation range of the process parameters of the chemical products and the correlation between the process parameters and the product indicators according to the process parameters in the chemical product database; The feature extraction data, the quality prediction results and the output prediction results of the chemical products in the current state are input into the recommendation system, and the recommended production parameters and operation strategies of the chemical products are obtained according to the data knowledge base of the chemical products, including: When the difference between the quality prediction result of the chemical product in the current state and the preset quality exceeds a first preset difference, generating corresponding recommended raw material component values ​​and recommended catalyst dosage according to the data knowledge base of the chemical product; When the difference between the output prediction result of the chemical product in the current state and the preset output exceeds a second preset difference, corresponding recommended process parameters are generated according to the data knowledge base of the chemical product.

2. The method for intelligently recommending chemical products based on data analysis according to claim 1, characterized in that: The method of using the chemical production prediction model to predict the feature extraction data to obtain the production prediction result of the chemical product in the current state includes: Inputting the feature extraction data into the chemical production prediction model; The chemical production prediction model calculates the production prediction value of the chemical product in the current state according to the process parameters in the feature extraction data; The output prediction value of the chemical product under the current state is normalized to obtain the output prediction result of the chemical product under the current state.

3. The method for intelligently recommending chemical products based on data analysis according to claim 2, characterized in that: The chemical production prediction model is a feedforward neural network model; The feedforward neural network model is pre-trained by the process parameters and output index values ​​in the chemical product database.

4. The method for intelligently recommending chemical products based on data analysis according to claim 1, characterized in that: The method of using the chemical quality prediction model to predict the feature extraction data to obtain the quality prediction result of the chemical product in the current state includes: Inputting the feature extraction data into the chemical quality prediction model; The chemical quality prediction model calculates the quality prediction value of the chemical product in the current state according to the raw material composition and catalyst usage in the feature extraction data; The quality prediction value of the chemical product in the current state is normalized to obtain the quality prediction result of the chemical product in the current state.

5. The method for intelligently recommending chemical products based on data analysis according to claim 4, characterized in that: The chemical quality prediction model is an XGBoost model. The XGBoost model is pre-trained by raw material composition, catalyst usage and quality index values ​​in the chemical product database.

6. The method for intelligently recommending chemical products based on data analysis according to claim 1, characterized in that: The method further comprises, Conduct safety monitoring of the process parameters and catalyst usage of the chemical products; When the temperature, pressure and reaction time in the process parameters of the chemical product exceed the corresponding preset ranges, a first safety warning is issued in the chemical production recommendation interface; When the catalyst usage in the catalyst usage of the chemical product exceeds a preset usage threshold, a second safety warning is issued on the chemical production recommendation interface.

Images