Aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis

Abstract

The invention discloses an aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis. The system comprises a sample cup used for containing an aluminum alloy melt sample; the temperature acquisition module is inserted into the sample cup, comprises a thermocouple assembly and is used for acquiring temperature data in a melt cooling process in real time; and the industrial control all-in-one machine is in communication connection with the temperature acquisition module, and aluminum alloy melt quality analysis and management software is arranged in the industrial control all-in-one machine. The software comprises a cooling curve processing function module, an intelligent criterion scoring function module, a database management function module and other function modules, is used for achieving temperature data preprocessing, feature value extraction, quality score calculation and result storage, and can output a diagnostic analysis result in combination with an interpretation algorithm. The method can be used for intelligently evaluating the quality of the aluminum alloy melt, has the advantages of quick response, high detection precision, traceable result, on-site deployment of the device and the like, and is suitable for quality control and production optimization of the melt in front of a furnace in the field of aluminum alloy casting.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aluminum alloy melt detection and quality control, in particular to an aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis. BACKGROUND

[0002] As a lightweight structural material, aluminum alloy is widely used in the fields of automobiles, aerospace, rail transportation and electronic devices, and the performance of its castings depends largely on the quality of the melt. Traditional melt quality control methods such as K-mold, PoDFA (Porous Disc Filtration Analysis) and metallographic / chemical composition analysis methods have certain accuracy, but they have significant problems such as poor real-time performance, dependence on experience, high equipment cost and non-traceability.

[0003] In recent years, thermal analysis as a simple, fast and low-cost online detection method has been widely used in the evaluation of aluminum alloy melt quality. This method reflects the solidification behavior of the melt by obtaining the temperature-time curve (i.e. cooling curve) during the cooling process. Traditional thermal analysis mainly relies on the characteristic points of these cooling curves (such as primary nucleation temperature and eutectic minimum undercooling temperature) for judgment, which can reflect the modification effect and crystallization behavior of the melt to some extent. However, this method still has the following significant shortcomings:

[0004] (1) Qualitative, lack of quantitative ability: Current thermal analysis in industry relies on experience interpretation or index comparison, which can only provide qualitative quality judgment and is difficult to achieve continuous variable scoring of melt quality;

[0005] (2) Strong dependence on experimental environment, poor repeatability: factors such as temperature collection accuracy, sample cup shape, environmental airflow and vibration can affect the stability of the cooling curve, resulting in large fluctuations and poor repeatability of the detection results;

[0006] (3) Single feature dimension, low information utilization rate: thermal analysis mostly extracts only the temperature values of the inflection points or platforms, and fails to effectively exploit the multi-dimensional thermodynamic / dynamic information contained in the cooling curve such as first derivative, second derivative, area and time interval, limiting its applicability and predictive ability in complex alloy systems.

[0007] In view of the above problems, there is an urgent need for an upgraded thermal analysis solution combining modern numerical algorithms, machine learning models and information platforms to further improve its quantitative accuracy, intelligent level and system integration capability in the evaluation of aluminum alloy melt quality while retaining its fast response and suitability for furnace detection. SUMMARY

[0008] In view of the problems of poor real-time performance, isolated data, lack of quantification and explainability of evaluation results in the existing aluminum alloy melt quality detection method, the purpose of the present application is to provide an aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis. The system realizes real-time acquisition of melt cooling curve through integrated hardware devices, completes data preprocessing, feature parameter extraction and intelligent scoring through special analysis software, and realizes long-term storage, visual management and diagnostic feedback of detection data in combination with structured database, thereby improving the intelligent, quantitative and standardized level of melt quality detection, and can be widely applied to furnace front quality control and production optimization of aluminum alloy casting site.

[0009] To achieve the above purpose, the present application adopts the following technical solutions:

[0010] An aluminum alloy melt quality detection system integrating database management and intelligent criterion analysis, comprising:

[0011] A sample cup for containing an aluminum alloy melt sample;

[0012] A thermocouple assembly inserted into the sample cup for real-time acquisition of temperature data during the melt cooling process;

[0013] A temperature acquisition module in communication connection with the thermocouple assembly for receiving and converting temperature signals;

[0014] An industrial computer in communication connection with the temperature acquisition module and installed with aluminum alloy melt quality analysis and management software;

[0015] The quality analysis and management software comprises:

[0016] A cooling curve processing module for smoothing filtering, derivative calculation and feature point extraction of temperature data;

[0017] An intelligent criterion analysis module for quality scoring of multi-dimensional cooling feature values based on a regression algorithm model and output of contribution of each feature to the scoring result in combination with an explanation algorithm;

[0018] A database management module for storing original temperature data, multi-dimensional feature values and scoring results.

[0019] In some embodiments, the system is integrated in a device box provided with a thermocouple insertion port and a temperature acquisition interface.

[0020] In some embodiments, the thermocouple assembly is a K-type or S-type thermocouple with a ceramic protection tube.

[0021] In some embodiments, the temperature acquisition module is a thermocouple input module with electrical isolation function and is connected to the industrial computer through a communication protocol.

[0022] In some embodiments, the cooling curve processing module performs Savitzky-Golay filtering smoothing, finite difference calculation of first and second order derivatives, and extraction of multi-dimensional feature parameters including nucleation temperature, minimum supercooling temperature, recalescence temperature rise, and cooling area.

[0023] In some embodiments, the intelligent criterion analysis module uses a random forest regression model and integrates a SHAP interpretation algorithm to achieve interpretability of the scoring results.

[0024] In some embodiments, the database management module includes a feature value data table (for recording feature values and scoring results), an original curve data table (for recording time-temperature-derivative information), and a log table and permission management module (for supporting access control, history recording, and data backup).

[0025] In some embodiments, the quality analysis and management software further includes a user interaction module that integrates an AI assistant function, supports natural language input, and generates diagnostic feedback in combination with the database and the scoring model.

[0026] In some embodiments, the device box is provided with an inclined sliding plate structure and contains a temperature insulation layer and a sliding rail buckle to improve the safety and stability of the equipment.

[0027] The application also provides an aluminum alloy melt quality detection method implemented by the above system, including the following steps:

[0028] Step one: pour the melt into the sample cup, and collect the cooling curve through the thermocouple assembly and the temperature acquisition module;

[0029] Step two: use the cooling curve processing module to pre-process the collected data and extract key feature parameters;

[0030] Step three: call the intelligent criterion analysis module to perform quality scoring on the feature parameters, and output the melt quality score and main influencing factor explanation;

[0031] Step four: store the feature values, scoring results, and original cooling curve through the database management module for subsequent query and traceability;

[0032] Step five: perform visual analysis through the user interaction module to view feature trends, historical comparisons, and obtain AI-assisted diagnostic suggestions.

[0033] Compared with the prior art, the application has the following beneficial effects:

[0034] (1) highly integrated device structure: realizes rapid deployment at the furnace front, improves detection efficiency and safety;

[0035] (2) Data full-process structured management: support for unified storage of cooling curves, characteristic values, score results and additional information, ensuring data traceability and reusability;

[0036] (3) Intelligent criterion analysis combined with interpretability: precise quantitative evaluation of melt quality based on artificial intelligence models, and decision basis provided through interpretation algorithms;

[0037] (4) Software and hardware integrated design: capable of interfacing with enterprise MES / ERP systems, suitable for online monitoring of melt quality and production optimization in intelligent foundry plants. BRIEF DESCRIPTION OF DRAWINGS

[0038] The accompanying drawings, which form a part of this disclosure, are intended to provide further understanding of the present disclosure and are incorporated herein for illustrative purposes. The schematic embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute undue limitations on the present disclosure. In the drawings:

[0039] Figure 1 is the overall roadmap of the system of the present disclosure.

[0040] Figure 2 is a schematic diagram of the melt sampling device of the embodiment of the present disclosure.

[0041] Figure 3 is a schematic diagram of the front structure of the industrial control all-in-one machine of the embodiment of the present disclosure.

[0042] Figure 4 is a schematic diagram of the interface panel of the industrial control all-in-one machine of the embodiment of the present disclosure.

[0043] Figure 5 is a schematic diagram of the workflow of the embodiment of the present disclosure.

[0044] Figure 6 is a database table structure diagram of the embodiment of the present disclosure.

[0045] Figure 7 is a schematic diagram of the interactive interface design of the embodiment of the present disclosure.

[0046] BRIEF DESCRIPTION OF DRAWINGS

[0047] 1. Sample cup, 2. Backflow device, 3. Overflow guide plate, 4. Thermocouple protection channel, 5. Anti-toppling device, 6. Industrial control all-in-one machine, 7. Industrial control machine power supply, 8. Audio interface, 9. Mouse and keyboard interface, 10. Video interface, 11. Network communication interface, 12. Thermocouple socket, 13. Temperature acquisition module. DETAILED DESCRIPTION

[0048] It should be noted that the embodiments of the present application and the features thereof can be combined with each other without conflict. The technical solutions of the present application will be described clearly and completely in combination with the drawings. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Other embodiments obtained by those skilled in the art based on the embodiments of the present application without creative labor also belong to the protection scope of the present application.

[0049] The aluminum alloy melt quality detection system of the embodiment of the present application comprises a sample cup 1, a thermocouple assembly, a temperature acquisition module and an industrial computer 6. The sample cup is used to hold an aluminum alloy melt sample; the thermocouple assembly is inserted into the sample cup 1 and used to acquire temperature data in real time during the cooling process of the melt; the temperature acquisition module is used to receive and convert the temperature signals acquired by the thermocouple assembly and send the converted signals to a computing terminal; the industrial computer 6 serves as the computing terminal and is installed with aluminum alloy melt quality analysis and management software.

[0050] The aluminum alloy melt quality analysis and management software comprises a cooling curve processing module, which is used to pre-process, calculate the derivative and extract feature points from the acquired temperature data; an artificial intelligence scoring module, which is based on a preset regression algorithm model to score the quality of multi-dimensional cooling characteristic values and supports outputting the contribution of each feature to the scoring result through a SHAP value explanation mechanism; a database management module, which comprises a structured characteristic value database and a cooling curve original data database and is used to store the feature data and scoring results of all detection batches; and a user interaction module, which supports data query, curve display, feature comparison, scoring result explanation and AI assistant man-machine dialogue functions. The above-mentioned system is integrally arranged in a device box with a protective structure, and the device box is provided with a thermocouple insertion port, a temperature acquisition interface, a display screen, an industrial keyboard and mouse port and a remote communication interface.

[0051] Referring to Figure 2 The melt sampling unit of the embodiment of the present application comprises a disposable sample cup 1, which is used to hold an aluminum alloy melt sample and is connected with a thermal analysis temperature measurement end. A backflow device 2 is arranged above and on the outside of the sample cup 1, which is used to guide the melt to flow back to a safe area when abnormal overflow occurs; an overflow guide plate 3 is arranged adjacent to the side surface of the sample cup 1, which is arranged obliquely relative to the sample cup 1 and is used to limit and guide the overflow direction of the melt to avoid high-temperature metal from falling into the electrical cavity of the device. In order to protect the temperature measurement wire and the interface, a thermocouple protection channel 4 is arranged between the sample cup 1 and the device box, and the thermocouple lead is arranged along the channel 4 and is electrically connected with the data acquisition and control unit at the rear end. Anti-toppling devices 5 are further arranged outside the device to limit and clamp the sample cup 1 and its supporting seat, so as to prevent the sample cup 1 from overturning during sampling and carrying, thereby improving the operation safety and ensuring the stability of data acquisition.

[0052] Referring toFigure 3 and Figure 4 The industrial computer 6 is a man-machine interface and a computing terminal with a capacitive touch display screen, and is used to run software modules such as cooling curve processing, artificial intelligence scoring, and database management. The interface panel of the industrial computer 6 is provided with an industrial computer power supply 7, an audio interface 8, a mouse and keyboard interface 9, a video interface 10, and a network communication interface 11 to meet the needs of power supply, voice prompt / alarm, external input device, external display, and wired / wireless communication, respectively. To realize rapid connection of the temperature measurement channel in front of the furnace, the interface panel is also provided with a thermocouple socket 12. The thermocouple socket 12 is electrically connected to a temperature acquisition module 13 through a shielded cable. The temperature acquisition module 13 is preferably a thermocouple input / conversion module with electrical isolation function, supports K-type or S-type thermocouple signal input, and communicates with the industrial computer 6 through an industrial bus such as Modbus or RS485, so as to realize real-time acquisition, transmission, and host computer processing of the temperature signal. The relative positional relationship and connection mode of the above components and interfaces are shown in Figure 3 and Figure 4 .

[0053] The cooling curve processing module performs the following on the original temperature data: (1) Savitzky-Golay filter smoothing processing; (2) finite difference calculation of first and second derivatives; and (3) extraction of multi-dimensional characteristic values including nucleation temperature, minimum supercooling temperature, recalescence temperature rise, and under-area.

[0054] The artificial intelligence scoring module uses a trained random forest regression model and uses the SHAP explanation algorithm to visually display the scoring influence of each feature. The database management module includes a characteristic value data table (records the sample number, time, furnace number, characteristic value, and scoring result of each detection), an original curve data table (records time-temperature-derivative and other columns of information, which is used for subsequent superimposed analysis and backtracking verification), and a log table and permission management module (supports historical change record, access control, and backup recovery).

[0055] The AI assistant module is based on an embedded generative large model framework, supports natural language question input, and generates suggestions or diagnostic opinions for melt quality evaluation in combination with database knowledge.

[0056] In a specific application, the thermocouple in the sample cup 1 is led out through the thermocouple protection channel 4 and connected to the thermocouple socket 12, and the signal conditioning and digitization processing is completed by the temperature acquisition module 13, and the collected data is transmitted to the industrial computer 6 through the network communication interface 11 or the local bus. When abnormal overflow occurs, the backflow device 2 and the overflow guide plate 3 cooperate to change the movement path of the liquid metal, and cooperate with the limiting action of the anti-toppling device 5 to reduce the risk of melt overflow to the equipment and personnel. The industrial computer power supply 7 provides an independent channel for system power supply, the audio interface 8 can be linked to issue an acoustic alarm, and the mouse and keyboard interface 9 and the video interface 10 facilitate the rapid deployment of external input devices and external display, so that the system of the application still has good maintainability and industrial adaptability in complex field environment.

[0057] The aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis provided by the application, the core of which is to realize automatic acquisition of melt temperature data through integrated hardware devices, to complete cooling curve feature recognition and AI intelligent scoring through special analysis software, and to establish a structured database platform to realize long-term storage and visual management of quality data. The system can be widely applied to online monitoring and analysis of melt quality in aluminum alloy casting sites.

[0058] The specific implementation process of the system and method includes the following steps:

[0059] S1: Melt sample loading and temperature data acquisition.

[0060] Sample is taken from an aluminum alloy smelting furnace, and the melt is injected into a disposable sample cup pre-equipped with a K-type or S-type thermocouple. The thermocouple acquires temperature signals through a temperature acquisition module with electrical isolation function and transmits them to an industrial touch control all-in-one machine in real time. The device box is provided with a thermocouple protection channel and an inclined overflow guide plate to ensure the safety of the sampling process and the stability of data acquisition.

[0061] S2: Cooling curve analysis and multi-dimensional feature extraction.

[0062] The industrial all-in-one machine runs the aluminum alloy melt quality analysis software to process the collected temperature data as follows:

[0063] Smooth filtering is performed on the original data (such as using the Savitzky-Golay method);

[0064] The first and second derivatives of the temperature curve are calculated to identify the inflection points and platforms in the cooling process;

[0065] Multiple characteristic values are automatically extracted, such as nucleation temperature, minimum supercooling temperature, recalescence temperature rise, cooling section area, derivative extreme value, and solidification time difference.

[0066] The extracted multi-dimensional features constitute the input data set for melt quality analysis.

[0067] S3: AI model analysis and quality scoring.

[0068] The pre-trained random forest regression model is used to evaluate the above feature data, outputting a melt quality score. The model integrates the SHAP (Shapley Additive Explanations) algorithm to explain the contribution of each feature to the score result, generating visual explanation charts to help operators understand the scoring basis. The scoring range can be set to 0-100 points, and the scoring results can be divided into excellent, good, medium, and poor levels.

[0069] S4: Data writing and database structured management.

[0070] The database management module of the system will write the following data during the detection process to the local database:

[0071] Sample basic information (number, time, furnace number, material type, etc.);

[0072] Cooling curve raw data (three columns of data: time, temperature, and first derivative);

[0073] All extracted feature values and their units;

[0074] Melt quality score results and corresponding SHAP weight analysis;

[0075] History of AI assistant interaction and generated suggestions.

[0076] The database uses structured table design and JSON extended fields, with functions such as permission management, historical data retrieval, and automatic backup.

[0077] S5: Visual analysis and intelligent human-computer interaction.

[0078] Operators can browse the current cooling curve, historical scoring trends, and feature changes through the graphical interface. The AI assistant module built into the system supports natural language queries, such as inputting "What is the reason for the low melt quality of this batch?" or "What is the difference from the previous furnace?", and the system will generate corresponding diagnostic text feedback based on the scoring model and database records.

[0079] Example: The specific application process of an aluminum alloy melt quality detection system is as follows (its device structure and composition are shown in Figure 2 , and the system workflow is shown in Figure 5 ):

[0080] 1. Start the detection system. Turn on the portable integrated device, including the industrial touch all-in-one machine, temperature acquisition module, and thermocouple positioning assembly; start the analysis software, complete system initialization, database self-checking, and equipment communication joint debugging.

[0081] 2. On-site sampling and real-time data acquisition. Take a spoonful of aluminum alloy melt from the smelting furnace, quickly pour it into a disposable sample cup (a K-type thermocouple is pre-embedded in the sample cup and connected to a temperature acquisition module of model ADAM-6018+), and the system starts recording the temperature-time data of the melt cooling in real time. The data acquisition frequency is set to 10 Hz, and the temperature recording stops when it drops to about 530°C.

[0082] 3. Cooling curve preprocessing and feature extraction. The analysis software automatically completes the denoising processing of the original temperature data (using Savitzky-Golay filtering), finite difference derivative calculation, and extracts the following main feature points: primary nucleation temperature, primary minimum undercooling temperature, primary recalescence temperature rise, primary crystal segment area, eutectic minimum undercooling temperature, eutectic segment area, solidification interval, etc., a total of 13 characteristic parameters.

[0083] 4. Intelligent criterion scoring analysis. The extracted characteristic values are input into the embedded random forest regression model (model training sample number > 80, determination coefficient R² > 0.85), and the quality score of the melt for this furnace (0-100 points) is obtained. At the same time, the SHAP method is used for visual analysis of the scoring basis, identifying the influence of each feature on the scoring result.

[0084] 5. Results written into the database. The database system constructed in this embodiment is used to support the data acquisition, feature extraction, model scoring, and historical tracing functions of the entire aluminum alloy melt quality detection process. The core of its design is to structurally manage multi-source heterogeneous data such as cooling process, chemical composition, and mechanical properties, ensuring the accuracy, traceability, and industrial adaptability of the data. Referring to Figure 6 , the database structure of the system includes four core sub-tables: cooling curve record table, cooling curve data table, alloy composition data table, and mechanical property data table, which are associated with each other through foreign keys, and centered on the main table cooling curve record table, realizing horizontal linkage and vertical traceability of data. Specifically as follows:

[0085] As shown in Figure 4 , the system database is composed of the following four core sub-tables, which are associated with each other through foreign keys, and centered on the main table cooling curve record table, realizing horizontal linkage and vertical traceability:

[0086] (1) Cooling curve record table (CoolingCurveRecords table)

[0087] This table is the main record table for all melt test batches, recording basic experimental information and main characteristic parameters. Each record is connected to other tables through a unique RecordID as the primary key, enabling data aggregation.

[0088] Field Name Data Type Meaning Description RecordID INT (Primary Key, Auto-increment) Cooling record number Device VARCHAR(20) Device number Date DATE Test date Time TIME Sampling time Operator VARCHAR(10) Operator ID ID VARCHAR(10) Sample ID maxT FLOAT Maximum temperature Liqui FLOAT Liquidus temperature TN FLOAT Primary nucleation temperature TU FLOAT Primary minimum undercooling temperature TG FLOAT Primary growth temperature TEN FLOAT Eutectic nucleation temperature TEU FLOAT Eutectic minimum undercooling temperature TEG FLOAT Eutectic growth temperature TF FLOAT Solidification end temperature tN FLOAT Primary nucleation time tU FLOAT Primary minimum undercooling time tG FLOAT Primary growth time tEN FLOAT Eutectic nucleation time tEU FLOAT Eutectic minimum undercooling time tEG FLOAT Eutectic growth time tF FLOAT Solidification end time ΔT1 FLOAT Primary recalescence temperature rise ΔT2 FLOAT Eutectic recalescence temperature rise ΔTEG FLOAT Eutectic growth temperature difference before and after modification Δt1 FLOAT Primary undercooling recalescence time Δt2 FLOAT Eutectic undercooling recalescence time Δt3 FLOAT Eutectic growth solidification end time difference Primary Area FLOAT Primary area: area enclosed by primary nucleation point, primary minimum undercooling temperature point, and solidification end temperature, cooling heat release in primary reaction section. Eutectic Area FLOAT Eutectic area: area enclosed by eutectic nucleation point, eutectic minimum undercooling temperature point, and solidification end temperature, cooling heat release in eutectic reaction section. AvgFirstDiv FLOAT First derivative average value, speed of cooling rate in primary and eutectic reaction sections MaxSecDiv FLOAT Second derivative maximum value, change speed of cooling rate in primary and eutectic reaction sections

[0089] This table serves as the core driving structure of the database, establishing a correlation relationship with all subsequent data tables through RecordID or CurveID, enabling unified control and quality recording of cooling curve data.

[0090] (2) Cooling curve data table (CoolingCurveRawData table)

[0091] Field Name Data Type Description RawDataID INT (Primary Key, Auto-increment) Unique identifier of raw data CurveID INT (Foreign Key) Corresponding RecordID in CoolingCurveRecords CoolingTime FLOAT Cooling time record Temp FLOAT Temperature record isDegassed BOOLEAN Whether degassing operation is performed, 1 for yes, 0 for no isSlagRemoved BOOLEAN Whether slag removal operation is performed, 1 for yes, 0 for no

[0092] The data comes from the temperature collection module. This table supports batch insertion, furnace-by-furnace calling, and abnormal curve filtering, assisting the system in derivative calculation and feature extraction.

[0093] (3) Alloy composition data table (AlloyComposition table)

[0094] Field Name Data Type Description CompositionID INT (Primary Key, Auto-increment) Unique identifier of alloy composition CurveID INT (Foreign Key) Corresponding RecordID in CoolingCurveRecords Date DATE Record date Composition_Si FLOAT Si content in aluminum alloy Composition_Mg FLOAT Mg content in aluminum alloy Composition_Ti FLOAT Ti content in aluminum alloy Composition_Sr FLOAT Sr content in aluminum alloy Composition_Fe FLOAT Fe content in aluminum alloy Composition_Mn FLOAT Mn content in aluminum alloy …

[0095] This table provides a quantitative basis for subsequent system research on the correlation between composition, curve, and performance, and can be expanded to support more than 10 elements.

[0096] (4) Mechanical properties data table (MechanicalProperties table)

[0097] Field Name Data Type Description PropertyID INT (Primary Key, Auto-Increment) Unique identifier of mechanical property CurveID INT (Foreign Key) Corresponding RecordID in CoolingCurveRecords YS FLOAT Yield strength (MPa) UTS FLOAT Tensile strength (MPa) EL FLOAT Elongation (%) Kmold FLOAT K mold method result Hydrogen FLOAT Hydrogen content in melt (mL / 100g)

[0098] 6、Data visualization and user interaction. In this embodiment, the system integrates a visual analysis module and an AI assistant interaction module on an industrial computer, which helps operators quickly obtain melt state information, analyze historical trends, and explain abnormal reasons in the production field or laboratory. The system mainly has the following functions:

[0099] (1) Cooling curve visualization function:

[0100] The system can draw the temperature-time cooling curve of the current sample and its corresponding first and second derivative curves; it can dynamically label key feature points (such as primary nucleation temperature TN, minimum undercooling temperature TU, solidification time, derivative extreme points, etc.) on the curve; it also supports displaying the scoring trend and key feature fluctuation of historical samples in the form of a broken line chart.

[0101] (2) Abnormal prompt and parameter deviation warning:

[0102] The system is embedded with a characteristic parameter threshold judgment function. Once a certain characteristic (for example, eutectic temperature rise, primary crystal area) exceeds the empirical range, a pop-up window warning is triggered. In addition, according to the change amplitude and trend of the score results, typical abnormal patterns such as "quality decline" and "stability fluctuation" can be identified, prompting the operator to carry out on-site treatment or furnace re-inspection. The system supports users to compare and analyze the current and historical cooling curves, score trends, and same batch data, and provides an abnormal reminder function.

[0103] (3) AI assistant interaction module:

[0104] The AI assistant module integrated by the system is a major technical highlight of the present application, which adopts a natural language human-computer interaction mode. Users can dialogue with the AI model through text input to obtain targeted quality analysis suggestions and data interpretation. The module supports the following two deployment modes: first, a local deployment edge large language model mode (LLM-on-Edge), that is, a simplified version of a large language model (such as MiniGPT, RWKV, DeepSeek, or a locally quantized Baichuan-13B model) is deployed locally on an industrial computer, combined with a local knowledge embedding vector library, to realize fast response and on-site application without the support of the external network; second, a network API calling mode. When the device has network conditions, the system can call cloud large model services such as OpenAI API, Ali Tongyi, and Baidu Wenxin Yiyang to obtain stronger language understanding and text generation capabilities.

[0105] (4) Interaction interface design:

[0106] As shown in Figure 7 , it is a schematic diagram of the interaction interface design in the embodiment of the present application. The visual interaction interface realizes a desktop GUI based on the Qt framework, and the interface design style is simple and friendly. The main functional areas include: a left navigation bar (providing function entrances such as sampling start, data management, historical query, and AI assistant), a middle cooling curve and derivative drawing area (supporting curve zooming in and out, and hovering to display coordinate values), a right analysis area (used to display model score results and abnormal prompt information), and a bottom status bar (displaying the current state of the system).

[0107] Compared with the prior art, the aluminum alloy melt quality detection system and method integrated with database management and intelligent criterion analysis of the present application have the following advantages: the device structure is highly integrated, can be quickly deployed on-site at the furnace, and significantly improves the detection efficiency and safety; the whole process data is managed in a structured manner, supporting quality historical traceability analysis and trend judgment; the artificial intelligence model scoring is introduced and combined with explainable analysis to realize precise quantitative evaluation and intelligent auxiliary decision of the melt quality; the software and hardware system is designed integrally, and can be connected to the enterprise MES / ERP system to meet the melt quality control requirements of intelligent foundry plants.

[0108] The application discloses an aluminum alloy melt quality detection system integrating database management and intelligent criterion analysis, which can collect temperature data in real time during the melt cooling process, automatically complete data preprocessing and multi-dimensional feature extraction, and score and interpret the melt quality by using an embedded intelligent criterion analysis model. The system is equipped with a structured database for storing cooling curve raw data, extracted characteristic values and scoring results, and supports visual management of melt quality data. The device has a compact overall structure, with protection and wire protection design, and is suitable for rapid installation and deployment on site. The application realizes intelligent, quantitative and standardized evaluation of aluminum alloy melt quality, has the advantages of on-site deployability, rapid response, high precision and strong traceability, and is suitable for furnace quality control and production optimization in the field of aluminum alloy casting.

[0109] The application proposes an aluminum alloy melt quality detection system and method integrating database management and intelligent criterion analysis, aiming at the problems of structural dispersion, data isolation, inability to quantify and lack of explainability of evaluation results in existing aluminum alloy melt thermal analysis methods. The system and method realize integrated design of the whole process from sampling, data acquisition and processing, feature recognition, model scoring, data storage to interactive feedback, greatly improving the intelligent and digital level of melt quality control. The application not only improves the efficiency and scientificity of melt quality evaluation, but also has good result explainability, data traceability and industrial adaptability, and is suitable for melt process control and quality digital standard construction under complex working conditions.

[0110] In the description of the present application, it should be understood that the terms "center", "longitudinal", "transverse", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer" and the like orientation or position relationship are based on the position relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, but not for limiting the scope of the present application.

[0111] In addition, the terms "first", "second" and the like are only used to distinguish different components or steps, and do not represent order or importance, nor are used to limit the number. Therefore, the features containing "first", "second" can explicitly or implicitly include one or more of the features. In this paper, the meaning of "multiple" is two or more, unless otherwise specified.

[0112] In the present application, unless otherwise explicitly specified, the terms "mounting", "connection", "linking", "fixing" and the like should be interpreted in a broad sense. For example, it can be fixed connection, or detachable connection, or the two as a whole; it can be mechanical connection, or electrical connection or communication connection; it can be directly connected, or indirectly connected through intermediate medium; it can be internal communication of elements, or interaction relationship between elements. The person skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0113] The above description is only the preferred embodiment of the present application, and is not intended to limit the present application. Any modification, equivalent replacement or improvement within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An aluminum alloy melt quality detection system integrating database management and intelligent criteria analysis, characterized in that, The system comprises: a sample cup for containing an aluminum alloy melt sample; a thermocouple assembly inserted into the sample cup for collecting temperature data in real time during the cooling process of the melt; a temperature acquisition module in communication with the thermocouple assembly for receiving and converting temperature signals; an industrial computer in communication with the temperature acquisition module and installed with aluminum alloy melt quality analysis and management software; wherein the quality analysis and management software comprises: a cooling curve processing module for pre-processing, derivative calculation and feature point extraction of temperature data; an intelligent criterion analysis module for quality scoring of multi-dimensional cooling characteristic values based on a regression algorithm model, and output of contribution degree explanations of each feature based on an interpretation algorithm; a database management module for storing original temperature data, multi-dimensional characteristic values and scoring results.

2. The system of claim 1, wherein, The system is integrated in a device box, which is provided with a thermocouple insertion port and a temperature acquisition interface.

3. The system of claim 2, wherein, The thermocouple assembly is a K-type or S-type thermocouple with a ceramic protection tube.

4. The system of claim 1, wherein, The temperature acquisition module is a thermocouple input module with electrical isolation function, and is connected to the industrial computer through a communication protocol.

5. The system of claim 1, wherein, The cooling curve processing module performs Savitzky-Golay filter smoothing processing, finite difference calculation of first and second derivatives, and automatic extraction of multi-dimensional characteristic parameters including nucleation temperature, minimum supercooling temperature, recalescence temperature rise and cooling area.

6. The system of claim 1, wherein, The intelligent criterion analysis module uses a random forest regression model and integrates a SHAP interpretation algorithm to output feature contribution degree analysis results.

7. The system of claim 1, wherein, The database management module includes a characteristic value data table for recording sample characteristic values and scoring results, an original curve data table for recording time-temperature-derivative information, and a log table and a permission management module for supporting access control, historical records and data backup.

8. The system of claim 1, wherein, The quality analysis and management software further comprises a user interaction module, which integrates an AI assistant function, supports natural language input and generates diagnostic feedback based on the database and the scoring model.

9. The system of claim 2, wherein, The device box is provided with an inclined sliding plate structure and contains a temperature insulation layer and a sliding rail buckle.

10. A method for detecting the quality of an aluminum alloy melt using the system according to any one of claims 1-5, characterized in that, The method comprises the following steps: Step 1: Pour the aluminum alloy melt into the sample cup, and obtain the temperature curve of the cooling process through the thermocouple assembly and the temperature acquisition module; Step 2: Pre-process the collected temperature data using the cooling curve processing module, and extract key characteristic parameters; Step 3: Call the intelligent criterion analysis module to score the extracted multi-dimensional characteristic values, and output the melt quality score results and main influencing factor explanations; Step 4: Store the characteristic values, scoring results and original cooling curves in the database through the database management module for subsequent query and traceability; Step 5: View the characteristic trend, historical data comparison and AI assistant-based diagnostic analysis suggestions through the user interaction module.

Images