Database management system based on building reinforcement material performance analysis

Abstract

The invention relates to the crossing field of building material science and information technology, and discloses a database management system based on building reinforcement material performance analysis, which comprises the following steps: constructing a causal path map which comprises a source node, a mechanism node and a performance node and forms a causal path representing an action mechanism; receiving experimental data, and performing forward deduction along a causal path by using a transfer function to obtain a performance prediction value; according to the deviation degree of the predicted value and the measured value, dynamically adjusting the confidence degree of each causal path by adopting a preset updating rule to realize the self-evolution of the atlas; and when experimental data cannot be explained by all high-confidence paths, mechanism abnormal points are judged, and a new hypothesis path is generated. According to the method, mechanism modeling and data driving are combined, white box processing of the analysis process and dynamic evolution of a knowledge system are achieved, the interpretability of the result is remarkably improved, knowledge gaps can be actively found, and scientific guidance is provided for material research and development.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of building materials science and information technology, specifically a database management system based on the performance analysis of building reinforcement materials. Background Technology

[0002] Building reinforcement materials are crucial for improving and repairing the performance of engineering structures. Their final macroscopic performance, such as mechanical strength and durability, is influenced by a complex coupling of multiple factors, including material composition, manufacturing process, and microstructural evolution. With the development of materials science, the research and development and quality control of these reinforcement materials increasingly rely on the management and in-depth analysis of large amounts of experimental data.

[0003] In existing technologies, the management and analysis of such experimental data typically employs a combination of traditional relational databases and statistical or machine learning models. Conventional database systems perform well in structured data storage and efficient retrieval, effectively archiving and querying completed experimental results. However, these systems are designed primarily for data recording, and their inherent tabular data models struggle to effectively characterize the complex, nonlinear, and inherently causal chains of physical or chemical logic between material composition, process parameters, intermediate structural features, and final properties. The internal mechanisms of action within materials are often implicit in the data and cannot be explicitly and structurally expressed as part of a data model.

[0004] To uncover deeper correlations within data, the industry has adopted various machine learning algorithms, such as neural networks and support vector machines, to construct predictive models of material properties. These data-driven methods can establish highly accurate input-output mappings under specific conditions, but the models themselves typically operate as black boxes. This means that while the models can predict which components will produce which properties, they cannot provide a clear, scientifically based physical explanation—that is, they cannot answer why such properties occur. This lack of interpretability constitutes a significant technical obstacle in scenarios requiring a deep understanding of the underlying mechanisms, such as new formulation design, process optimization, or failure mechanism diagnosis.

[0005] Furthermore, existing knowledge representation methods, even those that partially employ graph structures to describe relationships between entities, largely remain at a static and descriptive level. Once these systems are built, their internal knowledge is typically fixed, lacking an effective, automated mechanism to dynamically verify, correct, and even evolve the credibility of their stored knowledge or theories based on continuously inputting new experimental evidence. Therefore, when experimental results deviate from model predictions, the system cannot self-adjust, nor can it proactively and specifically reveal potential defects or blind spots within the current knowledge system. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a database management system based on the performance analysis of building reinforcement materials. This system solves the problem that existing database management systems can only handle the apparent relationships between data, but lack the structured expression and computational ability of the intrinsic mechanism of material performance, resulting in opaque analysis and prediction processes and a lack of scientific explanation.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a database management system based on the performance analysis of building reinforcement materials, comprising: A causal path graph construction module is configured to construct a causal path graph. The causal path graph is a data structure used to express the mechanism of material properties, comprising: at least one source node representing a directly set material composition or process parameter; at least one performance node representing a measurable macroscopic engineering property of the material; at least one mechanism node representing a physical or chemical process and serving as an intermediate node between the source node and the performance node; and at least one causal path, which is a sequence of nodes formed by sequentially connecting one source node, at least one mechanism node, and one performance node in a causal order.

[0008] The data processing module, which is connected to the causal path graph construction module, is configured to receive experimental data containing source node parameter values ​​and performance node measured values, and update the confidence of the causal path based on the experimental data.

[0009] The reasoning and analysis module, which is connected to the data processing module, is configured to analyze or display the causal path based on the updated confidence level.

[0010] In one possible implementation, the causal path graph further includes edges connecting the nodes, and each edge is associated with a transfer function that defines the numerical influence of the source node on the target node. In this case, the data processing module is specifically configured to: perform forward inference based on the source node parameter values ​​in the experimental data and using the transfer functions associated with each edge on the causal path to obtain the predicted values ​​of the performance nodes; and update the confidence level of the causal path based on the deviation between the predicted values ​​and the measured values ​​of the performance nodes.

[0011] Furthermore, the data processing module calculates the deviation degree corresponding to the deviation using the following formula. ; Wherein, (?) represents the deviation, a dimensionless parameter used to quantify the relative difference between the predicted and measured values; ) is the predicted value of the performance node; ) represents the measured value of the performance node.

[0012] Furthermore, the data processing module updates the confidence level of the causal path using the following formula: ; in,( () represents the updated confidence level; () represents the confidence level before the update; ) is a preset learning sensitivity factor, which is a constant greater than zero; (?) is the deviation calculated based on the newly received experimental data; It is a natural exponential function.

[0013] In another possible implementation, the data processing module is further configured to: identify the experimental data as a mechanistic anomaly when the calculated prediction biases of all high-confidence causal paths related to a certain experimental data are higher than a preset bias threshold; and generate at least one new hypothetical path in the causal path graph for the mechanistic anomaly. The method for generating the hypothetical path includes: establishing a new temporary edge between two existing nodes belonging to different paths to form the hypothetical path; and inserting a new mechanistic node between two consecutive nodes of an existing causal path to form the hypothetical path.

[0014] In another possible implementation, the causal path map further includes structural nodes, which represent quantifiable structural features of the material at the microscopic or mesoscopic scale. In this case, the causal path is formed by connecting the origin node, the mechanism node, the structural node, and the performance node in causal order.

[0015] In another possible implementation, the reasoning and analysis module is specifically configured to: return causal paths related to the user query; and visually distinguish and display the returned causal paths based on the current confidence level of each causal path. Furthermore, the reasoning and analysis module can also be configured to: analyze the confidence distribution of all causal paths in the causal path graph to identify and output mechanism nodes or causal paths with confidence levels below a preset knowledge threshold, as knowledge gaps.

[0016] A second aspect of this invention provides a database management method based on the performance analysis of building reinforcement materials, the method comprising the following steps: S1: Construct a causal path graph, which includes at least one source node, at least one performance node, at least one mechanism node as an intermediate node, and at least one causal path formed by sequentially connecting the nodes. S2: Receive experimental data containing the parameter values ​​of the source node and the measured values ​​of the performance node; S3: Update the confidence level of the causal path based on the experimental data; S4: Analyze or display the causal path based on the updated confidence level.

[0017] This invention provides a database management system based on the performance analysis of building reinforcement materials. It has the following beneficial effects: 1. This invention establishes mechanistic nodes representing physical or chemical processes and structural nodes representing intermediate structural states, thereby constructing a causal path connecting the origin node and the performance node. This enables the system to structurally and computationally express the complete mechanism of a material's action, from composition to performance. Compared to existing technologies that only focus on the correlation between input and output, the analytical results of this invention possess clear causal logic and physical meaning, achieving white-box analysis of the process, thus significantly improving the interpretability and scientific rigor of performance prediction and failure analysis results.

[0018] 2. This invention establishes a closed-loop mechanism for self-evolution of the knowledge model by setting up a data processing module to compare the predicted and measured values ​​of causal paths using received experimental data to calculate the deviation. Based on this deviation, a negative exponential decay model is used to dynamically update the confidence level of each causal path, enabling the system's internal knowledge system to self-correct and improve based on external, objective experimental evidence. The reliability of its analysis and reasoning dynamically improves with the accumulation of data, overcoming the limitations of the static and fixed nature of traditional knowledge base models.

[0019] 3. This invention, by incorporating mechanisms for anomaly identification and hypothesis generation, enables the system to proactively and structurally construct new hypothetical paths when faced with experimental data that cannot be reasonably explained by existing high-confidence causal paths. This can be achieved through methods such as path bridging or mechanism insertion. This allows the invention not only to apply existing knowledge but also to proactively identify gaps and deficiencies in the current knowledge system, providing researchers with clearly defined and experimentally verifiable new research directions, thereby guiding scientific exploration and accelerating materials innovation. Attached Figure Description

[0020] Figure 1 This is a system architecture diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example: Please see the appendix Figure 1 -Appendix Figure 2 This invention provides a database management system based on the performance analysis of building reinforcement materials, including: The causal path graph construction module is used to construct a causal path graph, which includes: At least one source node representing material composition or process parameters; At least one performance node representing the macroscopic engineering properties of the material; At least one mechanism node representing a physical or chemical process, serving as an intermediate node between the source node and the performance node; At least one causal path consisting of source nodes, mechanism nodes, and performance nodes connected in sequence; In this embodiment, the causal path graph construction module, as the basic data structure generation unit of the entire system, has the core function of transforming unstructured or semi-structured domain knowledge in the field of building reinforcement materials, consisting of expert experience, theoretical formulas, and standard specifications, into a structured data model that can be recognized, calculated, and reasoned about by a computer—namely, the causal path graph. The output of this module provides a unified, mechanism-based analysis framework for subsequent data processing and reasoning / analysis modules.

[0023] Specifically, the construction process of the causal path graph construction module includes defining and instantiating the basic constituent elements of the graph, namely nodes and edges, and combining them into a higher-level logical structure, namely causal path.

[0024] First, regarding node definition, the module will define the nodes in the graph ( The nodes are divided into four mutually exclusive categories, and the set of all nodes is ( This ensures the semantic clarity and uniqueness of the model.

[0025] The four node categories are: source node ( These nodes are defined as the starting point of a causal chain, representing initial variables that can be directly set or controlled during material design or preparation. Preferably, intrinsic nodes may include, but are not limited to, material composition parameters, such as the water-cement ratio. ), the dosage of specific additives; or process parameters, such as curing temperature ( The prestress level of the reinforcing material. Using these parameters as source nodes aims to provide a definite and traceable input source for all performance analyses.

[0026] Performance nodes ( These nodes are defined as the endpoints of a causal chain, representing the final macroscopic engineering properties of the material that can be measured through standardized experiments. Preferably, performance nodes may include mechanical performance indicators, such as 28-day compressive strength (…). ); or durability performance indicators, such as chloride ion permeability coefficient ( The purpose of using these macroscopic performance metrics as performance nodes is to provide clear, engineering-relevant targets for system analysis and prediction.

[0027] Mechanism node ( These nodes, a core feature of this embodiment, represent specific physical or chemical processes occurring during the formation of material properties. Mechanism nodes act as bridges connecting origin nodes and performance nodes, or other intermediate nodes, revealing the intrinsic mechanisms of interaction between "input" and "output." For example, in concrete, mechanism nodes can be concretized as "cement hydration," "pozzolanic reaction," or "interfacial transition zone strengthening effect." By introducing mechanism nodes, the system can visualize abstract causal relationships as a series of analyzable intermediate steps.

[0028] Structural nodes ( These structural nodes represent quantifiable intermediate structural states formed at the microscopic or mesoscopic scale after a material undergoes specific mechanistic processes. Structural nodes are typically direct products of mechanistic nodes and serve as key structural parameters influencing final performance. The existence of structural nodes provides measurable intermediate evidence for verifying and quantifying mechanistic processes.

[0029] In summary, the complete set of nodes can be represented as the union of its subsets: .

[0030] Secondly, regarding the definition of edges and transfer functions, the causal path graph construction module uses directed edges ( ) to represent from the source node ( ) to target node ( The direct causal relationship of the graph. To achieve the computability of the graph, a key technical feature of this embodiment is that for each edge ( Force association with a transfer function ( ).

[0031] The function of a transfer function is to define and quantify the value of the source node in mathematical form. How does this affect or calculate the value of the target node? Their relationship can be formally expressed as: ; To ensure the system's broad applicability, the transfer function ( The specific form of the formula is flexible. It can be a recognized physical or chemical model in the field, such as Fick's second law describing the diffusion process of ions in porous media; it can also be an empirical or semi-empirical formula based on fitting a large amount of experimental data, such as a theoretical formula describing the relationship between concrete strength and porosity; in some qualitative relationships, it can even be a logical rule based on expert knowledge. Through this design, the system transforms qualitative causal relationships into quantitative numerical transmission.

[0032] Finally, regarding the combination and attribute assignment of causal paths, the module defines a causal path as a complete and uninterrupted ordered sequence of nodes and edges in the graph that starts from any source node, passes through at least one intermediate node, and finally reaches a performance node. .

[0033] ,in , Each causal path logically represents a complete and self-consistent theory of the material's performance mechanism. To characterize the credibility of these theories at the initial stage of their construction, the module also assigns a parameter to each causal path. Assign an initial confidence level The confidence level is a value between 0 and 1, and its initial value can be set by domain experts based on prior knowledge such as the degree of acceptance of the theory represented by the path in the current academic or engineering community and the degree of literature support.

[0034] By executing the above steps, the causal path graph construction module ultimately generates a structured, quantified causal path graph with an initial state. This graph not only graphically presents the intrinsic mechanisms of material properties, but more importantly, it provides a necessary, computable underlying data model for subsequent modules to perform advanced analytical functions such as evidence-based confidence evolution and path-based causal reasoning. This transforms static domain knowledge into a dynamic and evolvable foundation for system intelligence.

[0035] The data processing module is used to receive experimental data containing the parameter values ​​of the source node and the measured values ​​of the performance node, and update the confidence of the causal path based on the experimental data. In this embodiment, the data processing module is the core execution unit that connects to the causal path graph construction module and provides dynamic data to the reasoning and analysis modules. Its main function is to transform externally input, discrete experimental data into a basis for verifying or correcting the internal logical structure of the causal path graph. Through a series of preset calculation processes, it realizes the dynamic evolution of the graph confidence level and has the ability to identify knowledge gaps and generate new hypotheses under specific conditions.

[0036] Specifically, the functionality of the data processing module relies on a rigorous, multi-step internal workflow.

[0037] First, in the data reception and correlation phase, the module is configured to receive experimental data organized in a specific format. Preferably, each experimental data ( It contains two parts of information: a set of source node parameter values ​​corresponding to the source nodes in the graph ( ), and a set of measured performance node values ​​corresponding to the performance nodes ( ).

[0038] ; Upon receiving the experimental data, the module performs an evidence anchoring operation. This operation retrieves all causal pathways by searching the entire causal path graph, identifying pathways whose start and end points are of the same type as the current experimental data. This involves identifying causal paths that match the parameter types contained in the data, thus forming a set of paths to be verified that are directly related to the experimental data. The purpose of this step is to precisely link an isolated piece of external evidence to all the theoretical models that could possibly explain it.

[0039] Secondly, in the core stage of model evolution, the module targets each causal path in the aforementioned set of paths to be verified. The confidence dynamic algorithm is then initiated. This algorithm is a key technical feature of this embodiment, enabling the spectrum to self-correct based on external evidence.

[0040] The first step of the algorithm is forward path deduction. In this step, the module will use the experimental data ( ) measured values ​​of source node parameters ( ) as the initial input, and along the causal path ( ) node sequence Call each edge segment sequentially, level by level. The transfer function associated with ) Through the composition of functions, the theoretical predicted value of the performance node corresponding to the causal path is finally calculated. ).

[0041] ; This deduction process is essentially a complete numerical simulation of the physical or chemical mechanism represented by the path.

[0042] The second step of the algorithm is to measure the prediction bias. After obtaining the theoretical prediction value, the module compares it with the actual measured values ​​of the performance nodes in the experimental data. This is compared to quantify the gap between the theoretical model and objective facts. This gap is calculated by measuring a normalized, dimensionless bias ( ). ) is used to characterize it.

[0043] ; Deviation ( The value of ) directly reflects the causal path ( ) Regarding the current experimental data ( The accuracy of the interpretation. A lower deviation value means that the mechanism represented by the path fits the experimental results well.

[0044] The third step of the algorithm is the confidence update rule. This step is the core of realizing the map's "learning" ability. The module uses an update formula based on a negative exponential decay model, based on the bias calculated in the previous step (…). To adjust the causal path ( ) confidence level ( ).

[0045] ; In this formula, ( ) is a path ( The new confidence level after processing the current data; ) is its confidence level before the update; ) is a pre-defined, positive-zero learning sensitivity factor used to regulate the weight of the influence of single-experiment evidence on confidence updates; It is a natural exponential function. This rule ensures that when the predicted path deviates significantly from the actual path, its confidence level is subject to a corresponding degree of decay penalty, thereby dynamically tending the overall confidence distribution of the map to a state consistent with a large amount of objective experimental evidence.

[0046] In addition, the data processing module includes an anomaly-driven hypothesis generation mechanism for knowledge discovery. This mechanism is designed to handle unexpected experimental results that cannot be explained by existing knowledge systems.

[0047] In this mechanism, the module first identifies mechanistic anomalies. When a piece of experimental data ( After input, if all related values ​​have a confidence level higher than a certain preset confidence threshold, then... The causal path of ) and the calculated deviation ( All of them are higher than another preset deviation threshold. When ), the data point ( This condition is considered an anomaly in the mechanism. This criterion means that none of the theories currently considered credible in the system can provide a reasonable explanation for this experimental phenomenon.

[0048] Once an anomaly in the mechanism is identified, the module will automatically initiate a hypothesis generation process. Preferably, this process can construct a new hypothesis path representing the unknown mechanism using one or a combination of the following two strategies: One strategy is path bridging, which involves establishing a temporary, new edge between two previously unconnected nodes in a causal path graph, thus forming a completely new causal path. This aims to explore potential, undiscovered synergistic or antagonistic effects between different parameters or intermediate processes. Another strategy is mechanistic insertion, which involves inserting a completely new, abstract mechanistic node between two consecutive nodes in an existing high-confidence path. This move is intended to indicate that existing theories may have overlooked a key intermediate reaction or process.

[0049] The newly generated hypothesis paths will be assigned an extremely low initial confidence level and marked by the system as pending verification. In this way, the data processing module can not only use data to quantitatively evaluate and revise existing knowledge, but also proactively and structurally propose new scientific questions, providing a clearly directional direction for materials science research.

[0050] The reasoning and analysis module is used to analyze or display causal paths based on the updated confidence level.

[0051] In this embodiment, the reasoning and analysis module, as the system's final output and human-computer interaction interface, primarily transforms the causal path graph established by the causal path graph construction module and dynamically updated by the data processing module into visual and quantifiable analytical results that provide direct guidance to the user. The effective operation of this module relies on a structured knowledge model with confidence levels, provided by the aforementioned modules and validated and corrected by experimental data.

[0052] Specifically, the core function of the reasoning and analysis module is reflected in its execution of diverse causal reasoning and knowledge mining tasks.

[0053] On one hand, the module provides a weighted causal query function. When it receives a user-initiated query request regarding the relationship between a specific source node and a performance node, the module first traverses the entire causal path graph to retrieve all causal paths connecting the two endpoint nodes being queried. Unlike traditional database queries that only return a list of associations, the key feature of this module is that it assigns a weighted causal confidence level to each retrieved causal path. This means that the values ​​are repeatedly updated and iterated by the data processing module, and then analyzed.

[0054] Subsequently, the module generates a visual weighted causal network graph using these paths and their confidence information. Preferably, in this visual graph, the visual attributes of different causal paths, such as line thickness, color saturation, or transparency, are correlated with the confidence value of that path (…). This involves direct correlation. For example, paths with higher confidence levels may have thicker lines and more prominent colors; conversely, paths with lower confidence levels are represented by thinner, darker lines. In this way, users can intuitively distinguish which mechanisms are supported by extensive experimental data and are highly reliable dominant pathways, while others are still in the theoretical speculation stage, with weak evidence support, or are highly controversial. This function transforms complex, multi-dimensional spectral data into a clear, easy-to-understand, and hierarchical knowledge structure for users.

[0055] On the other hand, the module also features a proactive knowledge gap identification function. This function does not require the user to specify a particular query; instead, the module performs periodic or reactive global scanning and analysis of the entire causal path graph. During this process, the module aims to identify knowledge gaps within the graph.

[0056] Preferably, the criteria for identifying a node or path as a knowledge gap are that it simultaneously meets two conditions: first, the node or path has a significant structural impact on one or more key performance nodes in the topology of the graph; for example, it is a necessary path to a certain key performance node, or it participates in forming multiple causal paths affecting that performance; second, the confidence level of the node or path itself (…). If a node or path meets the above conditions, it is identified as a knowledge gap. The module marks such gaps as knowledge gaps and generates an analysis report for the user. This report clearly identifies which key mechanisms within the current knowledge system have questionable reliability and require experimental verification. This function provides researchers with precise data guidance for planning future experimental directions and optimizing the allocation of research and development resources.

[0057] Furthermore, to fully utilize the computability of the causal path graph, the module can also perform forward simulation. When the user inputs one or more sets of hypothetical source node parameter values, the module will invoke path retrieval and calculation logic similar to weighted causal queries, following the high-confidence causal path related to the input parameters, and utilizing the transfer functions associated with each edge of the path (…). It performs forward calculations to predict the possible values ​​or trends of relevant performance nodes under the given assumptions. The output can be a single predicted value based on the path with the highest confidence, or an expected value range or probability distribution obtained by comprehensively considering multiple paths and their confidence weights.

[0058] Correspondingly, the module can also perform reverse causal tracing to assist in failure analysis. When the user inputs a set of known initial material parameters (source node values) and an unexpected performance result that has occurred (e.g., strength far below design values), the module can traverse the spectrum in reverse. By analyzing the paths that start from the failure performance node and point backward to each intermediate mechanism node and structural node, and combining the confidence level of each path with the sensitivity of the transfer function, the module can calculate and highlight the failure mechanism node or deteriorated structural node most likely to cause the performance anomaly.

[0059] In summary, the reasoning and analysis module, through the combination of the above functions, transforms a dynamically evolving data structure into an interactive analysis tool capable of responding to queries, proactively discovering, simulating predictions, and tracing diagnoses. This realizes the transformation from data storage to knowledge services, providing in-depth, mechanism-based decision support for the performance analysis and innovative research and development of building reinforcement materials.

[0060] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A database management system based on the performance analysis of building reinforcement materials, characterized in that, The system includes: A causal path graph construction module is used to construct a causal path graph, which includes: At least one source node representing material composition or process parameters; At least one performance node representing the macroscopic engineering properties of the material; At least one mechanism node representing a physical or chemical process, the mechanism node serving as an intermediate node between the source node and the performance node; At least one causal path consisting of the source node, the mechanism node, and the performance node connected in sequence; The data processing module is used to receive experimental data containing the parameter values ​​of the source node and the measured values ​​of the performance node, and update the confidence of the causal path based on the experimental data. The reasoning and analysis module is used to analyze or display the causal path based on the updated confidence level.

2. The system according to claim 1, characterized in that, The causal path graph also includes edges connecting nodes, and the edges are associated with transfer functions used to quantify the causal effects between nodes; The data processing module is specifically used for: Based on the source node parameter values ​​in the experimental data, and using the transfer function on the causal path to perform forward deduction, the predicted value of the performance node is obtained. The confidence level of the causal path is updated based on the deviation between the predicted value and the measured value of the performance node.

3. The system according to claim 2, characterized in that, The data processing module calculates the deviation degree corresponding to the deviation using the following formula: ； in,( ) represents the deviation degree, ( ) is the predicted value of the performance node, ( ) represents the measured value of the performance node.

4. The system according to claim 3, characterized in that, The data processing module updates the confidence level of the causal path using the following formula: ； in,( ) represents the updated confidence level, ( () represents the confidence level before the update. ) is a preset learning sensitivity factor, ( The deviation is denoted as ). It is a natural exponential function.

5. The system according to claim 1, characterized in that, The data processing module is also used for: When the prediction bias of all high-confidence causal paths related to a certain experimental data exceeds a preset bias threshold, the experimental data is identified as a mechanism outlier. For the aforementioned mechanism anomalies, at least one new hypothetical path is generated in the causal path graph.

6. The system according to claim 5, characterized in that, The methods for generating at least one new hypothesis path include: A new temporary edge is created between two existing nodes belonging to different paths to form the hypothetical path; A new mechanistic node is inserted between two consecutive nodes in an existing causal path to form the hypothetical path.

7. The system according to claim 1, characterized in that, The causal path map also includes structural nodes representing the microscopic or mesoscopic structural features of the material; The causal path is formed by sequentially connecting the source node, the mechanism node, the structure node, and the performance node.

8. The system according to claim 1, characterized in that, The reasoning and analysis module is specifically used for: Based on the user's query, return the causal path related to the query; Based on the current confidence level of each causal path, the returned causal paths are visually distinguished and displayed.

9. The system according to claim 1, characterized in that, The reasoning and analysis module is also used for: The confidence distribution of all causal paths in the causal path graph is analyzed to identify and output mechanism nodes or causal paths with confidence levels below a preset knowledge threshold, as knowledge gaps.

10. A database management method based on the performance analysis of building reinforcement materials, wherein the database management system based on the performance analysis of building reinforcement materials according to any one of claims 1-9 is characterized in that, Includes the following steps: S1: Construct a causal path graph, which includes at least one source node, at least one performance node, at least one mechanism node as an intermediate node, and at least one causal path formed by sequentially connecting the nodes. S2: Receive experimental data containing the parameter values ​​of the source node and the measured values ​​of the performance node; S3: Update the confidence level of the causal path based on the experimental data; S4: Analyze or display the causal path based on the updated confidence level.

Images