Iron modified biochar environmental risk and remediation benefit evaluation system and method

Abstract

The application discloses an iron modified biochar environmental risk and remediation benefit evaluation system and method, relates to the field of environmental remediation data evaluation, and can adjust semantic similarity threshold value when it is detected that current search performance does not reach a preset standard, and synchronously optimizes cache capacity configuration when necessary, so that optimal parameter combinations can be found within a limited number of adjustments, thereby avoiding the problem of low search efficiency caused by improper parameter setting in traditional search; this not only greatly reduces invalid search time, but also improves cache resource utilization, and guarantees the comprehensiveness and accuracy of search results; finally, the search results are sorted through a sorting model, so that the most effective remediation impact features can be identified and applied in priority, thereby guiding the accurate preparation and release of iron modified biochar, and significantly improving the remediation effect of contaminated sites; at the same time, through a regular environmental risk evaluation mechanism, the safety and sustainability of the remediation process are ensured.

Description

Technical Field

[0001] This invention belongs to the field of environmental remediation data assessment, and specifically relates to a system and method for assessing the environmental risks and remediation benefits of iron-modified biochar. Background Technology

[0002] Iron-modified biochar, as an emerging functional composite material, combines the structural characteristics of biochar with the high redox activity of iron, and also has excellent catalytic performance and environmental stability. In recent years, it has received widespread attention in the fields of wastewater treatment and soil and water environment remediation. In the process of searching for the impact characteristics of iron-modified biochar remediation at contaminated sites, traditional search methods often generate a large number of invalid search operations due to the lack of intelligent parameter optimization mechanisms, resulting in a huge waste of computing resources. At the same time, it is difficult to guarantee the comprehensiveness and reliability of search results. In the practice of contaminated site remediation, this inefficient search method directly affects the accurate identification and effective application of remediation impact characteristics, resulting in a lack of precise guidance for the application of iron-modified biochar and difficulty in guaranteeing the remediation effect. Summary of the Invention

[0003] To address the problems in related technologies, this invention proposes an environmental risk and remediation benefit assessment system and method for iron-modified biochar, in order to overcome the aforementioned technical problems existing in the current related technologies.

[0004] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution: This invention relates to a method for assessing the environmental risks and remediation benefits of iron-modified biochar, comprising the following steps: S1. Select several groups of historical iron-modified biochar remediation site remediation process effects and influence characteristics, and label and rank them. S2. Based on the labeled and sorted data in S1, construct the final sorted feature mapping model to repair the impact. S3. Obtain historical data on the cache capacity, semantic similarity threshold, cache hit rate, and retrieval time of several sets of environmental remediation impact feature retrieval data at various levels, and construct a final mapping model of cache hit rate and retrieval time for remediation impact feature retrieval. S4. Input the cache capacity data at each level corresponding to the current repair impact feature retrieval process and the semantic similarity threshold into the S3 mapping model for mapping; S5. Compare the mapping result in S4 with the corresponding preset threshold. If the requirements are met, no adjustment is needed. Otherwise, adjust the current semantic similarity threshold repeatedly based on the mapping model in S3, and map the adjusted semantic similarity threshold again. If the mapping result meets the requirements within the limit number of times, the adjustment is complete. Otherwise, proceed to S6. S6. Repeatedly adjust the retrieval cache capacity, and combine it with the semantic similarity threshold adjusted in S5 to input into the mapping model in S3 for mapping until the mapping result meets the requirements; perform the retrieval operation based on the finally determined cache capacity and similarity threshold, sort the retrieval results through the mapping model in S2, and then prepare and apply iron-modified biochar, and conduct environmental risk and remediation benefit assessment.

[0005] Preferably, step S1 includes the following steps: S11. Several characteristic types affecting the environmental remediation effect of iron-modified biochar on contaminated sites are defined to obtain a set of biochar remediation influence characteristic types. This set includes raw material type, raw material particle size, pyrolysis temperature during preparation, type of modifier during preparation, modifier concentration during preparation, adsorption equilibrium time, specific surface area, optimal adsorption pH, and dosage during environmental remediation. Then, several characteristic indicators reflecting the environmental remediation effect of iron-modified biochar are defined to obtain a set of biochar remediation effect characteristic types. This set includes measured adsorption capacity, heavy metal passivation rate, and pH adjustment capability. S12. Based on the set of contaminated site feature types, the set of biochar remediation impact features, and the set of biochar remediation effect features, collect biochar remediation impact feature data and biochar remediation effect feature data corresponding to several historical instances of using iron-modified biochar for environmental remediation of contaminated sites, and obtain historical biochar remediation impact feature dataset and historical biochar remediation effect feature dataset. S13. Using the historical biochar remediation effect feature dataset, the remediation impact feature data of each group in the historical biochar remediation impact feature dataset are labeled to obtain the historical labeled biochar remediation impact feature dataset; the historical labeled biochar remediation impact feature dataset is randomly grouped to obtain the historical labeled remediation impact feature data set; Based on the historical biochar remediation effect feature dataset, and by manually sorting the data in each feature data group of the historical labeled remediation impact feature data group according to the effect quality and removing the labeled data, the historical sorted remediation impact feature data group is obtained. By implementing a strategy that combines random sampling with manual sorting, the objectivity and impartiality of the data analysis phase are ensured, while expert experience and knowledge are fully integrated to improve the reliability of the results. Furthermore, it provides training samples for the construction of a subsequent hybrid retrieval and sorting model, thereby enabling a more accurate assessment of various environmental remediation effectiveness metrics.

[0006] Preferably, step S2 includes the following steps: S21. Based on the historical post-annotation repair impact feature data set and the historical sorting repair impact feature data set, construct a mapping model with the input of the post-annotation repair impact feature data set and the output of the sorting repair impact feature data set, and obtain the final sorting repair impact feature mapping model. By establishing a mapping model, we can not only effectively explore the contribution of each influencing factor to the remediation effect, but also quickly predict the relative advantages and disadvantages of the remediation effect when facing new remediation scenarios by inputting the corresponding biochar preparation and application parameters, thus providing a scientific basis for the optimization of remediation schemes. In turn, it can provide efficient and accurate decision support for the environmental remediation of contaminated sites, significantly improving the scientific nature and efficiency of remediation work.

[0007] Preferably, step S3 includes the following steps: S31. During the process of searching for the impact features of iron-modified biochar remediation of contaminated sites, a multi-level search system cache is set to obtain a remediation impact feature search cache set. In conjunction with the remediation impact feature search cache set, the total capacity data, search semantic similarity threshold, search cache hit rate data, and search time data of the corresponding levels of contaminated site remediation impact feature search cache are obtained in several historical searches for the impact features of iron-modified biochar remediation of contaminated sites. This yields a historical search cache capacity dataset, a historical search semantic similarity threshold set, a historical search cache hit rate dataset, and a historical search time dataset. S32. Based on the historical retrieval cache capacity dataset, historical retrieval semantic similarity threshold dataset, historical retrieval cache hit rate dataset, and historical retrieval time dataset, construct a mapping model between retrieval cache capacity data, retrieval semantic similarity threshold, retrieval cache hit rate data, and retrieval time data to obtain the final repair impact feature retrieval cache hit rate time mapping model. The mapping model effectively links the retrieval semantic similarity threshold with caching efficiency, enabling the system to optimize retrieval speed by dynamically adjusting the threshold while ensuring retrieval accuracy. In addition, the model can also identify the impact patterns of different retrieval accuracies on caching performance, providing targeted guidance for system optimization.

[0008] Preferably, step S4 includes the following steps: S41. Label the data in the current iron-modified biochar environmental remediation impact feature knowledge base with remediation effect feature data; after the labeling is completed, obtain the total capacity data of the environmental remediation impact feature retrieval cache at each level during the current retrieval process and the corresponding preset retrieval semantic similarity threshold according to the remediation impact feature retrieval cache set, and obtain the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold. S42. Input the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model to obtain the current retrieval cache hit rate dataset and the current retrieval time data; Simultaneously, a hybrid retrieval strategy is used to obtain multi-level cache capacity data and preset thresholds to construct a dynamic retrieval dataset, providing a basis for subsequent dynamic adjustments to cache capacity and semantic similarity thresholds, balancing efficiency and accuracy while reducing redundant calculations; it also provides a data basis for the subsequent elastic allocation of system resources.

[0009] Preferably, step S5 includes the following steps: S51. Based on the characteristics of the impact of repair, retrieve the cache set and the current retrieval requirements, set the hit rate threshold of each level of cache in the current retrieval process and the current retrieval time threshold to obtain the current retrieval cache hit rate threshold set. S52. Set a first adjustment number threshold; if there is a cache hit rate data in the current retrieval cache hit rate dataset that is less than or equal to the hit rate threshold corresponding to the current retrieval cache hit rate threshold set, or if the current retrieval time data is greater than or equal to the current retrieval time threshold, the current retrieval semantic similarity threshold is repeatedly adjusted to obtain the current adjusted retrieval semantic similarity threshold. Otherwise, no adjustment is needed; the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold. S53. The current adjusted retrieval semantic similarity threshold is combined with the current retrieval cache capacity dataset and input into the final repair feature retrieval cache hit rate time mapping model to obtain the current adjusted retrieval cache hit rate dataset and the current adjusted retrieval time data. If the number of repetitions in S52 is less than or equal to the first adjustment number threshold, and there is a cache hit rate in the current adjusted cache hit rate dataset that is greater than the hit rate threshold corresponding to the current cache hit rate threshold set, and the current adjusted retrieval time data is less than the current retrieval time threshold, the adjustment is complete, and the current retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; otherwise, proceed to S6. When the cache hit rate is found to be below standard or the retrieval time exceeds expectations, the retrieval efficiency is improved by iteratively optimizing the semantic similarity threshold. The number of adjustments can be reasonably controlled according to the preset adjustment threshold, which not only ensures that the system maintains stable performance under different data scales and retrieval complexities, but also avoids performance fluctuations caused by excessive adjustments.

[0010] Preferably, step S6 includes the following steps: S61. Repeatedly adjust the current retrieval cache capacity dataset to obtain the current adjusted retrieval cache capacity dataset; input the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model for mapping to obtain the current second-adjusted retrieval cache hit rate dataset and the current second-adjusted retrieval time data. The adjustment is complete when there is a cache hit rate in the current second-adjusted retrieval cache hit rate dataset that is greater than the hit rate threshold corresponding to the current retrieval cache hit rate threshold set and the current second-adjusted retrieval time data is less than the current retrieval time threshold; the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold. S62. Adjust the cache capacity and semantic similarity threshold at each level during the current retrieval of the knowledge base data on the environmental remediation impact of iron-modified biochar based on the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; perform the retrieval operation after the adjustment is completed. After the retrieval in S63 and S62 is completed, the current environmental remediation impact feature retrieval dataset is obtained; the current environmental remediation impact feature retrieval dataset is input into the final sorted remediation impact feature mapping model for mapping, and the current sorted environmental remediation impact feature retrieval dataset is obtained. Iron-modified biochar was prepared from the top-ranked data in the current sorted environmental remediation impact feature retrieval dataset. After preparation, it was applied to the corresponding contaminated site. After application, the environmental risk and remediation effect of the contaminated site were periodically assessed. When the initial search parameters fail to meet performance requirements, the cache capacity configuration can be automatically optimized and adjusted repeatedly. In addition, the search effect is continuously improved through the mapping model feedback mechanism to ensure that the optimal balance is achieved in the two key indicators of hit rate and response time, thereby ensuring the stability and efficiency of the search process.

[0011] An environmental risk and remediation benefit assessment system for iron-modified biochar includes a historical remediation impact feature data annotation and sorting module, a remediation impact feature mapping model construction module after sorting, a remediation impact feature retrieval cache hit rate time-consuming mapping model construction module, a current retrieval cache hit rate time-consuming mapping module, a current retrieval semantic similarity threshold adjustment module, and a current retrieval cache capacity determination and adjustment module.

[0012] The present invention has the following beneficial effects: 1. In this invention, when the current retrieval performance is detected to be below the preset standard, the semantic similarity threshold can be adjusted, and the cache capacity configuration can be optimized simultaneously when necessary. This ensures that the optimal parameter combination is found within a limited number of adjustments, thereby avoiding the problem of low retrieval efficiency caused by improper parameter settings in traditional retrieval. This not only significantly reduces invalid retrieval time and improves the utilization rate of cache resources, but also ensures the comprehensiveness and accuracy of retrieval results. Finally, the retrieval results are sorted by a ranking model, so that the most effective remediation impact features can be identified and applied first, thereby guiding the precise preparation and application of iron-modified biochar, thus significantly improving the remediation effect of contaminated sites. At the same time, the safety and sustainability of the remediation process are ensured through a regular environmental risk assessment mechanism.

[0013] 2. This invention can reasonably control the number of adjustments according to a preset threshold, avoiding performance fluctuations caused by excessive adjustments, and ultimately achieving the optimal match between the retrieval cache capacity and the semantic similarity threshold, which significantly improves the retrieval efficiency and user experience of the iron-modified biochar environmental remediation knowledge base.

[0014] 3. In this invention, when the initial retrieval parameters cannot meet the performance requirements, the cache capacity configuration can be automatically and repeatedly optimized and adjusted; in addition, the retrieval effect is continuously improved through the mapping model feedback mechanism to ensure that the optimal balance is achieved in the two key indicators of hit rate and response time; therefore, not only is the adaptive capability of the system improved, but also the stability and efficiency of the retrieval process are guaranteed through continuous performance monitoring and parameter tuning.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the environmental risk and remediation benefit assessment method for iron-modified biochar according to the present invention. Figure 2 This is a schematic diagram illustrating the process of constructing the final sorting and repairing the feature mapping model according to the present invention. Figure 3 This is a flowchart illustrating the process of constructing the final repair time-consuming mapping model for the feature retrieval cache hit rate in this invention. Figure 4 This is a schematic diagram illustrating the process of mapping the current retrieval cache hit rate and time consumption according to the present invention; Figure 5 This is a schematic diagram illustrating the process of adjusting the current retrieval semantic similarity threshold according to the present invention; Figure 6 This is a schematic diagram of the modules of an iron-modified biochar environmental risk and remediation benefit assessment system of the present invention; Figure 7 This is a schematic diagram illustrating the process of determining and adjusting the current retrieval cache capacity according to the present invention. Detailed Implementation

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0019] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0020] Example 1 Please see Figure 1 This embodiment describes a method for assessing the environmental risks and remediation benefits of iron-modified biochar. The specific preparation steps for iron-modified biochar are as follows: First, waste pomelo peels were ground and sieved to obtain pomelo peel biomass powder with an average particle size of 175 μm. Then, the obtained pomelo peel biomass powder was mixed with a 1.0 mol / L ferric chloride aqueous solution and reacted. The solid-liquid ratio of pomelo peel biomass powder to ferric chloride solution was 10 g / 150 mL, the reaction temperature was 70 °C, and the reaction time was 60 min. After the reaction was completed, the reaction solution was separated into solid and liquid phases to obtain iron-loaded biomass. Finally, the iron-loaded biomass was pyrolyzed at a temperature of 580 °C for 150 min to obtain iron-modified biochar. The method includes the following steps: S1. Select several groups of historical iron-modified biochar remediation site remediation process effects and influence characteristics, and label and rank them. Please see Figure 2 S1 includes the following steps: S11. Define several characteristic types that affect the environmental remediation effect of iron-modified biochar on contaminated sites, and obtain a set of biochar remediation influence characteristic types. The set of biochar remediation influence characteristic types includes: raw material type (e.g., wheat straw, coconut shell, robust scab, medicinal residue, manure, etc.), raw material particle size (e.g., 100–200 mesh, affecting impregnation uniformity and pyrolysis efficiency), pyrolysis temperature during preparation (e.g., 450–850℃), type of modifier during preparation (e.g., FeCl3, nano-zero-valent iron, etc.), modifier concentration during preparation (mol / L), adsorption equilibrium time (min), and specific surface area (m²). 2 The optimal adsorption pH and dosage (g / L) during environmental remediation are determined; several characteristic indicators reflecting the environmental remediation effect of iron-modified biochar are then set to obtain a set of biochar remediation effect characteristic types; the set of biochar remediation effect characteristic types includes the measured adsorption amount (mg / g), heavy metal passivation rate, and pH adjustment ability (monitoring the change of soil pH after biochar addition to evaluate its improvement effect on acidic or alkaline soils; positive numbers can be used to measure this, and the larger the value, the stronger the pH adjustment ability). S12. Based on the set of contaminated site feature types, the set of biochar remediation impact features, and the set of biochar remediation effect features, collect biochar remediation impact feature data and biochar remediation effect feature data corresponding to several historical instances of using iron-modified biochar for environmental remediation of contaminated sites, and obtain historical biochar remediation impact feature dataset and historical biochar remediation effect feature dataset. Partial data from the historical biochar remediation impact feature dataset and the historical biochar remediation effect feature dataset are shown in Table 1 and Table 2 below, respectively: Table 1. Examples of partial data on the impact characteristics of biochar remediation. Types of raw materials Average particle size of raw materials (mesh) Pyrolysis temperature (°C) Modifier type Modifier concentration (mol / L) Adsorption equilibrium time (min) Specific surface area (m² / g) Optimal adsorption pH Dosage (g / L) wheat straw 150 600 <![CDATA[FeCl3]]> 0.5 120 250 6.2 2.0 Coconut shell 200 700 Nano-zero valent iron 0.3 180 320 6.0 1.5 Strong bristle algae 150 500 <![CDATA[FeCl3]]> 0.8 90 180 5.8 2.5 dregs 250 800 Nano-zero valent iron 0.6 150 280 6.5 1.8 manure 200 650 <![CDATA[FeCl3]]> 0.4 100 220 6.1 2.2 Table 2. Examples of partial data on the characteristics of biochar remediation effects. Measured adsorption capacity (mg / g) Heavy metal passivation rate (%) pH regulation ability 120 78.5 0.8 150 85.2 1.2 95 72.1 0.6 135 81.3 1.0 110 75.6 0.9 S13. Using the historical biochar remediation effect feature dataset, the remediation impact feature data of each group in the historical biochar remediation impact feature dataset are labeled to obtain the historical labeled biochar remediation impact feature dataset; the historical labeled biochar remediation impact feature dataset is randomly grouped to obtain the historical labeled remediation impact feature data set; Based on the historical biochar remediation effect feature dataset, and by manually sorting the data in each feature data group of the historical labeled remediation impact feature data group according to the effect quality and removing the labeled data, the historical sorted remediation impact feature data group is obtained. By systematically constructing an evaluation system for the remediation effect of iron-modified biochar, data-driven, standardized, and intelligent management of the environmental pollution control process has been achieved. Firstly, starting from the source, the key factors affecting remediation effectiveness were identified. Through refined classification across multiple dimensions, including raw material type, particle size control, pyrolysis process parameters, and the selection and concentration adjustment of modifiers, scientific guidance is provided for the customized production of subsequent remediation materials. Simultaneously, a comprehensive evaluation index system was established, encompassing measured adsorption capacity, heavy metal passivation rate, and pH adjustment capability. This overcomes the limitations of traditional single-indicator evaluations, thus comprehensively reflecting the effects of biochar in practical application scenarios. The system has multiple functions; secondly, based on the accumulation of historical restoration case data, a large-scale sample database with time series characteristics is formed, and each record is given a clear quality label by the data annotation mechanism, which enhances the learning accuracy and generalization ability of the training model; on this basis, the operation strategy of combining random sampling and manual sorting is implemented, which not only ensures the objectivity and fairness of the data analysis stage, but also fully integrates expert experience and knowledge to improve the reliability of the results. Among them, the benchmark sequence is formed by manually sorting the historical dataset, which provides training samples for the construction of the subsequent mixed retrieval and sorting model, and can more accurately weigh the various environmental restoration effect measurement indicators. S2. Based on the labeled and sorted data in S1, construct the final sorted feature mapping model to repair the impact. S2 includes the following steps: S21. Based on the historical post-annotation repair impact feature data set and the historical sorting repair impact feature data set, construct a mapping model with the input of the post-annotation repair impact feature data set and the output of the sorting repair impact feature data set, and obtain the final sorting repair impact feature mapping model. S21 includes the following steps: S211. Construct an initial sorting and repair impact feature mapping model and set a first training data ratio (e.g., 8:2 or 7:3, which can be adjusted adaptively according to the actual training situation); divide the historical labeled repair impact feature data set and the historical sorting repair impact feature data set according to the first training data ratio to obtain the first training dataset and the first test dataset respectively. S212. Set a first training error threshold (10%~15%, which can be adjusted adaptively according to the actual training situation); input the first training dataset into the initial sorting and repairing influence feature mapping model for training; during the training process, if the training error is less than the first training error threshold, stop training and obtain the trained sorting and repairing influence feature mapping model; otherwise, continue training until the training error is less than the first training error threshold. S213. Set a first test accuracy threshold (90%~95%, which can be adjusted adaptively according to the actual test situation); input the first test dataset into the trained sorted post-repair influence feature mapping model for testing; after the test is completed, obtain the first test accuracy data; if the first test accuracy data is greater than or equal to the first test accuracy threshold, use the trained sorted post-repair influence feature mapping model as the final sorted post-repair influence feature mapping model; otherwise, return to S212 to continue training the trained sorted post-repair influence feature mapping model and repeat S213 until the first test accuracy data is greater than or equal to the first test accuracy threshold. The initial sorting followed by feature mapping repair uses a deep neural network (DNN) model, the structure of which is shown below: Input layer: Receives multi-dimensional metadata (such as raw material type, raw material particle size, pyrolysis temperature during preparation, optimal adsorption pH, and dosage during environmental remediation) of the historical annotation and repair impact feature data group. The input dimensions are determined according to the number of features. Fully connected layers: Layer 1: 512 neurons, using ReLU (Rectified Linear Function) activation function, with batch normalization introduced to accelerate training; Layer 2: 256 neurons, ReLU activation function; Layer 3: 128 neurons, ReLU activation function; Output layer: 1 neuron, using the Sigmoid activation function, outputs the ranking score (0-1 interval). Optimizer: Adam optimizer, with a learning rate of 0.001 and a loss function of mean squared error or cross-entropy; A deep neural network (DNN) model is used to construct a post-ranking repair impact feature mapping model. This model can automatically learn the complex mapping relationship between labeled repair impact feature data and manually ranked sequences through its multi-layer nonlinear processing units. The DNN model has powerful feature learning capabilities, enabling end-to-end mapping from raw metadata (such as evidence level, authority, and timeliness) to the final ranking result, eliminating the tedious manual feature engineering process. Specifically, the DNN model receives multi-dimensional metadata through the input layer, performs nonlinear transformations through multiple hidden layers (such as fully connected layers), enhances the network's expressive power using activation functions (such as ReLU), and finally generates ranking scores through the output layer. This structure allows the model to capture the deep, nonlinear relationship between metadata and ranking results, thereby achieving more accurate ranking results that better meet the needs of environmental repair in tasks such as repair impact feature retrieval. By training on historical remediation data, this model uses various characteristic parameters affecting remediation effectiveness (such as raw material type, pyrolysis temperature, modifier type and concentration, etc.) as input variables and the ranking of remediation effectiveness as the output target, thus establishing a nonlinear relationship between characteristic parameters and remediation effectiveness. This mapping relationship not only effectively mines the contribution of each influencing factor to the remediation effect but also, when facing new remediation scenarios, can quickly predict the relative merits of remediation effectiveness by inputting corresponding biochar preparation and application parameters, providing a scientific basis for optimizing remediation schemes. Furthermore, the model construction process fully utilizes the annotation and ranking information in historical data, enabling the model to learn not only the relationship between a single feature and its effect but also the comprehensive ranking rules of remediation effectiveness under multiple feature combinations, thereby improving the model's generalization ability and prediction accuracy. The final mapping model possesses good interpretability and practicality, providing efficient and accurate decision support for environmental remediation of contaminated sites, significantly improving the scientific rigor and efficiency of remediation work. S3. Obtain historical data on the cache capacity, semantic similarity threshold, cache hit rate, and retrieval time of several sets of environmental remediation impact feature retrieval data at various levels, and construct a final mapping model of cache hit rate and retrieval time for remediation impact feature retrieval. Please see Figure 3 S3 includes the following steps: S31. During the process of searching for the impact features of iron-modified biochar remediation of contaminated sites, a multi-level search system cache is set to obtain a remediation impact feature search cache set. In conjunction with the remediation impact feature search cache set, the total capacity data, search semantic similarity threshold, search cache hit rate data, and search time data of the corresponding levels of contaminated site remediation impact feature search cache are obtained in several historical searches for the impact features of iron-modified biochar remediation of contaminated sites. This yields a historical search cache capacity dataset, a historical search semantic similarity threshold set, a historical search cache hit rate dataset, and a historical search time dataset. For example, consider data from a contaminated site remediation impact feature retrieval system: The system's caching architecture comprises three levels: L1 fast cache, L2 capacity cache, and L3 persistent cache. The L1 cache, located in memory with a capacity of 2GB, stores frequently accessed repair feature data, such as common raw material types (wheat straw, coconut shells, etc.) and standard pyrolysis temperature ranges (450-850℃). The L2 cache, deployed on SSDs with a capacity of 10GB, stores historical repair case data accessed at moderate frequencies, including detailed modifier concentration ratios (0.3-0.8 mol / L) and specific surface area measurements (180-320 m² / g), among other moderately popular information. The L3 cache, built on mechanical hard drives with a capacity of 50GB, is used for archiving complete repair history records and cold data. During historical retrieval, the system recorded performance metrics for different caching levels: when the semantic similarity threshold was set to 0.85, the L1 cache hit rate reached 92%, with an average retrieval time of only 15 milliseconds; when the similarity threshold was adjusted to 0.75, the L2 cache hit rate was 78%, and the retrieval time increased to 45 milliseconds; and when the retrieval involved complex historical case comparisons, the similarity threshold was reduced to 0.65, requiring access to the L3 cache, with a hit rate of 65% and a retrieval time of 120 milliseconds. Analysis of the historical retrieval cache capacity dataset revealed that the optimal capacity allocation for the L1 cache is 15% of the total retrieval data, the L2 cache should hold approximately 40% of the data, and the remaining 45% can be archived to the L3 cache. S32. Based on the historical retrieval cache capacity dataset, historical retrieval semantic similarity threshold dataset, historical retrieval cache hit rate dataset, and historical retrieval time dataset, construct a mapping model between retrieval cache capacity data, retrieval semantic similarity threshold, retrieval cache hit rate data, and retrieval time data to obtain the final repair impact feature retrieval cache hit rate time mapping model. S32 includes the following steps: S321. Construct an initial repair impact feature retrieval cache hit rate time-consuming mapping model and set a second training data ratio (e.g., 8:2 or 7:3, which can be adjusted adaptively according to the actual training situation); divide the historical retrieval cache capacity dataset, historical retrieval semantic similarity threshold dataset, historical retrieval cache hit rate dataset and historical retrieval time-consuming dataset according to the second training data ratio to obtain the second training dataset and the second test dataset respectively. S322. Set a second training error threshold (10% to 15%, which can be adjusted adaptively according to the actual training situation); input the second training dataset into the initial repair impact feature retrieval cache hit rate time-consuming mapping model for training; during the training process, if the training error is less than the second training error threshold, stop training and obtain the trained repair impact feature retrieval cache hit rate time-consuming mapping model; otherwise, continue training until the training error is less than the second training error threshold. S323. Set a second test accuracy threshold (90%~95%, which can be adjusted adaptively according to the actual test situation); input the second test dataset into the trained repair impact feature retrieval cache hit rate time-consuming mapping model for testing; after the test is completed, obtain the second test accuracy data; if the second test accuracy data is greater than or equal to the second test accuracy threshold, use the trained repair impact feature retrieval cache hit rate time-consuming mapping model as the final repair impact feature retrieval cache hit rate time-consuming mapping model; otherwise, return to S322 to continue training the trained repair impact feature retrieval cache hit rate time-consuming mapping model and repeat S323 until the second test accuracy data is greater than or equal to the second test accuracy threshold; The initial repair impact feature retrieval cache hit rate time-consuming mapping model adopts a multilayer perceptron model, and its specific structure example is as follows: Input layer: 4 neurons (corresponding to 4 features: cache capacity, semantic similarity threshold, hit rate, and time consumption); Hidden layers: First hidden layer: 64 neurons, using the ReLU activation function (to solve the gradient vanishing problem and accelerate convergence); Second hidden layer: 32 neurons, using the Leaky ReLU activation function (to alleviate the neuron "death" problem); Output layer: 1 neuron (predicts hit rate or time taken), using either the Sigmoid activation function (outputs probability value, suitable for hit rate prediction) or the linear activation function (directly regresses the time taken value). Key parameter settings: Batch size: 32 (balancing training speed and memory usage); Learning rate: 0.001 (initial value, dynamically adjusted with the Adam optimizer); Loss function: Mean squared error (MSE, used for time-consuming regression) or binary cross-entropy (used for hit rate classification); Regularization: L2 regularization (weight decay coefficient 0.01, to prevent overfitting); MLPs can fully capture the complex interactions between features through fully connected layers, making them suitable for handling and repairing nonlinear mapping relationships that affect cache capacity, thresholds, and performance in feature retrieval. The ReLU series of activation functions improves training efficiency and avoids the gradient saturation problem of traditional Sigmoid. In addition, since the input is structured numerical data, there is no need for local feature extraction (convolutional layers) or dimensionality reduction (pooling layers). By constructing a mapping model for the cache hit rate and time consumption of environmental remediation impact feature retrieval, the overall performance of the environmental remediation impact feature retrieval system can be significantly improved. Specifically, based on historical operational data collection from a multi-level caching architecture, the correlation between cache capacity configuration at each level and retrieval performance can be accurately grasped. By analyzing the dynamic relationship between cache hit rate and response time, a scientific basis for system capacity planning is provided. The mapping model effectively links the retrieval semantic similarity threshold with cache efficiency, enabling the system to optimize retrieval speed by dynamically adjusting the threshold while ensuring retrieval accuracy. In addition, the model can also identify the impact patterns of different retrieval precisions (i.e., semantic similarity thresholds) on cache performance, providing targeted guidance for system optimization. By continuously learning historical retrieval patterns, the model can predict performance under different cache configurations, providing decision support for system expansion and parameter tuning. S4. Input the cache capacity data at each level corresponding to the current repair impact feature retrieval process and the semantic similarity threshold into the S3 mapping model for mapping; Please see Figure 4 S4 includes the following steps: S41. Label the data in the current iron-modified biochar environmental remediation impact feature knowledge base with remediation effect feature data; after the labeling is completed, obtain the total capacity data of the environmental remediation impact feature retrieval cache at each level and the corresponding preset retrieval semantic similarity threshold according to the remediation impact feature retrieval cache set during the current retrieval process (using a hybrid retrieval strategy, combining keyword matching and semantic similarity calculation to retrieve candidate document sets from the iron-modified biochar environmental remediation impact feature knowledge base), and obtain the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold; S42. Input the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model to obtain the current retrieval cache hit rate dataset and the current retrieval time data; This approach optimizes the retrieval performance of environmental remediation impact features through data annotation and dynamic caching analysis. First, the knowledge base data is annotated with remediation effect feature data to ensure the accuracy of retrieval semantics. Simultaneously, a hybrid retrieval strategy (keyword matching and semantic similarity calculation) is used to obtain multi-level cache capacity data and preset thresholds to construct a dynamic retrieval dataset. A mapping model is employed to transform capacity and hit rate data into time consumption predictions. Annotation of remediation effect feature data improves the semantic interpretability of retrieved documents, avoiding ambiguity issues inherent in traditional keyword matching. The hybrid retrieval strategy balances efficiency and accuracy, providing a basis for subsequent dynamic adjustments to cache capacity and semantic similarity thresholds, and reducing redundant computation. The mapping model enables real-time time consumption prediction, providing data for subsequent elastic allocation of system resources. In summary, this solution significantly reduces the response latency of environmental remediation impact feature retrieval while enhancing the stability of complex queries. S5. Compare the mapping result in S4 with the corresponding preset threshold. If the requirements are met, no adjustment is needed. Otherwise, adjust the current semantic similarity threshold repeatedly based on the mapping model in S3, and map the adjusted semantic similarity threshold again. If the mapping result meets the requirements within the limit number of times, the adjustment is complete. Otherwise, proceed to S6. Please see Figure 5 S5 includes the following steps: S51. Based on the characteristics of the repair impact, retrieve the cache set and the current retrieval requirements, set the hit rate threshold of each level of cache and the current retrieval time threshold (which can be adaptively set according to actual retrieval requirements) to obtain the current retrieval cache hit rate threshold set. S52. Set a first adjustment number threshold; if there is a cache hit rate data in the current retrieval cache hit rate dataset that is less than or equal to the hit rate threshold corresponding to the current retrieval cache hit rate threshold set, or if the current retrieval time data is greater than or equal to the current retrieval time threshold, the current retrieval semantic similarity threshold is repeatedly adjusted to obtain the current adjusted retrieval semantic similarity threshold. Otherwise, no adjustment is needed; the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold. S53. The current adjusted retrieval semantic similarity threshold is combined with the current retrieval cache capacity dataset and input into the final repair feature retrieval cache hit rate time mapping model to obtain the current adjusted retrieval cache hit rate dataset and the current adjusted retrieval time data. If the number of repetitions in S52 is less than or equal to the first adjustment number threshold, and there is a cache hit rate in the current adjusted cache hit rate dataset that is greater than the hit rate threshold corresponding to the current cache hit rate threshold set, and the current adjusted retrieval time data is less than the current retrieval time threshold, the adjustment is complete, and the current retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; otherwise, proceed to S6. By establishing a dynamic threshold adjustment mechanism and iterative optimization process, an intelligent adaptive optimization system for retrieving the impact characteristics of iron-modified biochar remediation at contaminated sites was achieved. Firstly, the system can automatically set and adjust the hit rate thresholds and retrieval time thresholds at each level of the cache based on real-time retrieval needs and cache performance, forming a closed-loop performance monitoring and adjustment system. When the cache hit rate is found to be substandard or the retrieval time exceeds expectations, the system activates a dynamic adjustment mechanism for the semantic similarity threshold, repeatedly iterating to improve retrieval efficiency. This adaptive adjustment strategy not only ensures stable performance under different data scales and retrieval complexities but also achieves continuous optimization of the retrieval strategy through the feedback mechanism of the mapping model. The entire adjustment process possesses intelligent judgment capabilities, reasonably controlling the number of adjustments according to preset thresholds to avoid performance fluctuations caused by excessive adjustments. Ultimately, it achieves the optimal match between the retrieval cache capacity and the semantic similarity threshold, significantly improving the retrieval efficiency and user experience of the iron-modified biochar environmental remediation knowledge base. S6. Repeatedly adjust the retrieval cache capacity, and combine it with the semantic similarity threshold adjusted in S5 to input into the mapping model in S3 for mapping until the mapping result meets the requirements; perform retrieval operation based on the finally determined cache capacity and similarity threshold, sort the retrieval results through the mapping model in S2, and then prepare and apply iron-modified biochar, and conduct environmental risk and remediation benefit assessment. Please see Figure 7 S6 includes the following steps: S61. Repeatedly adjust the current retrieval cache capacity dataset to obtain the current adjusted retrieval cache capacity dataset; input the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model for mapping to obtain the current second-adjusted retrieval cache hit rate dataset and the current second-adjusted retrieval time data. The adjustment is complete when there is a cache hit rate in the current second-adjusted retrieval cache hit rate dataset that is greater than the hit rate threshold corresponding to the current retrieval cache hit rate threshold set and the current second-adjusted retrieval time data is less than the current retrieval time threshold; the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold. S62. Adjust the cache capacity and semantic similarity threshold at each level during the current retrieval of the knowledge base data on the environmental remediation impact of iron-modified biochar based on the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; perform the retrieval operation after the adjustment is completed. After the retrieval in S63 and S62 is completed, the current environmental remediation impact feature retrieval dataset is obtained; the current environmental remediation impact feature retrieval dataset is input into the final sorted remediation impact feature mapping model for mapping, and the current sorted environmental remediation impact feature retrieval dataset is obtained. Iron-modified biochar was prepared from the top-ranked data in the current sorted environmental remediation impact feature retrieval dataset. After preparation, it was applied to the corresponding contaminated site. After application, the environmental risk and remediation effect of the contaminated site were periodically assessed. For example, in a contaminated site remediation impact characteristic retrieval system, the system adopts a three-level cache structure of L1, L2, and L3. The L1 level cache capacity is set at 2000kB to store frequently accessed remediation characteristic data, such as common raw material types like wheat straw and coconut shells, and core parameters such as the standard pyrolysis temperature range of 450-850℃. The L2 level cache capacity is 10000kB to store historical remediation case data with medium frequency of access, including detailed modifier concentration ratios of 0.3-0.8mol / L and specific surface area measurements of 180-320m² / g, etc. The L3 level cache capacity is 50000kB to archive and store complete remediation history records and cold data. During the specific retrieval process, the system first annotates 5,000 data points in the current iron-modified biochar environmental remediation impact feature knowledge base with remediation effect features. The annotation content includes key parameters such as measured adsorption capacity, heavy metal passivation rate, and pH adjustment capability. Through a hybrid retrieval strategy, combining keyword matching and semantic similarity calculation, candidate document sets are retrieved from the knowledge base, with keyword matching weighted at 0.4 and semantic similarity calculation weighted at 0.6. During the retrieval process, the system obtains the total capacity data of each level of cache in real time: L1 level cache capacity is 2,000 kB, L2 level cache capacity is 10,000 kB, and L3 level cache capacity is 50,000 kB. At the same time, the preset retrieval semantic similarity threshold is 0.85. The cache capacity data and semantic similarity threshold are input into the final repair feature retrieval cache hit rate time mapping model for mapping calculation, and the current retrieval cache hit rate dataset is obtained, in which the L1 level cache hit rate is 0.92, the L2 level cache hit rate is 0.78, and the L3 level cache hit rate is 0.65. At the same time, the corresponding retrieval time data are obtained, which are 15ms, 45ms, and 120ms respectively. Based on the retrieval cache set affected by the repair and the current retrieval requirements, the system sets the cache hit rate thresholds for each level to 0.90, 0.75, and 0.60, and the retrieval time thresholds to 20ms, 50ms, and 150ms. When the system detects that the current retrieval cache hit rate is less than or equal to the corresponding threshold or the retrieval time is greater than or equal to the corresponding threshold, the system repeatedly adjusts the retrieval semantic similarity threshold. After the initial adjustment, the retrieval semantic similarity threshold is set to 0.80. After multiple iterations, the final adjusted retrieval semantic similarity threshold is determined to be 0.75. By combining the adjusted semantic similarity threshold with the current retrieval cache capacity dataset and inputting it into the mapping model, the current retrieval cache hit rate dataset after the first adjustment is obtained. The L1 cache hit rate is improved to 0.95, the L2 cache hit rate to 0.82, and the L3 cache hit rate to 0.70. Simultaneously, the retrieval time is reduced to 12ms, 38ms, and 100ms respectively. After multiple rounds of adjustment and optimization, the final retrieval cache capacity dataset is determined to be 2200kB for L1, 11000kB for L2, and 55000kB for L3, with a final retrieval semantic similarity threshold of 0.75. Based on the final determined cache capacity and semantic similarity threshold, the current retrieval process is adjusted, and the system executes the retrieval operation to obtain the current environmental remediation impact feature retrieval dataset, containing 150 candidate remediation schemes. The retrieved data was input into the final ranked environmental remediation impact feature mapping model for mapping, resulting in the current ranked environmental remediation impact feature retrieval dataset. The top five data points were: biochar prepared from wheat straw (adsorption capacity 120 mg / g, heavy metal passivation rate 78.5%, pH adjustment capacity 0.8), biochar prepared from coconut shell (adsorption capacity 150 mg / g, heavy metal passivation rate 85.2%, pH adjustment capacity 1.2), and biochar prepared from *Cladosporium spp.* (adsorption capacity 95 mg / g, heavy metal passivation rate 72.1%, pH adjustment capacity 0.6), etc. Based on the top-ranked remediation scheme data, iron-modified biochar was prepared and applied to the corresponding contaminated sites. After application, the environmental risk and remediation effect of the contaminated sites were regularly assessed to ensure the scientific nature and effectiveness of the remediation work. By establishing a multi-layered adaptive adjustment mechanism and a closed-loop optimization process, the intelligent dynamic optimization of the contaminated site remediation impact feature retrieval system is achieved. Through an iterative adjustment strategy, when initial retrieval parameters fail to meet performance requirements, the cache capacity configuration can be automatically and repeatedly optimized. Furthermore, a mapping model feedback mechanism continuously improves retrieval performance, ensuring an optimal balance between the two key indicators of hit rate and response time. This dual optimization mechanism not only enhances the system's adaptability but also ensures the stability and efficiency of the retrieval process through continuous performance monitoring and parameter tuning. The entire adjustment process forms a complete closed-loop control system. From parameter setting, dynamic adjustment, performance mapping to final retrieval execution, each link possesses intelligent judgment and optimization capabilities, enabling the system to automatically optimize cache strategies and retrieval parameters based on actual operating results, significantly improving the retrieval efficiency and accuracy of the environmental remediation knowledge base. Simultaneously, this solution establishes a complete chain from retrieval to remediation effect evaluation, realizing intelligent management of the entire process from data retrieval and scheme optimization to actual remediation, providing strong technical support for the precise remediation of contaminated sites.

[0021] Example 2 Please see Figure 6 This embodiment discloses an environmental risk and remediation benefit assessment system for iron-modified biochar. The system can implement the method of the above embodiment, including a historical remediation impact feature data annotation and sorting module, a remediation impact feature mapping model construction module after sorting, a remediation impact feature retrieval cache hit rate time-consuming mapping model construction module, a current retrieval cache hit rate time-consuming mapping module, a current retrieval semantic similarity threshold adjustment module, and a current retrieval cache capacity determination and adjustment module. The historical remediation impact feature data annotation and sorting module is used to select several groups of historical impact features of iron-modified biochar remediation of contaminated sites and to annotate and sort them. The sorted post-repair impact feature mapping model construction module is used to construct the final sorted post-repair impact feature mapping model based on the annotation and sorted data in the historical repair impact feature data annotation and sorting module. The module for constructing the repair impact feature retrieval cache hit rate time mapping model is used to obtain several sets of historical data on the retrieval cache capacity, retrieval semantic similarity threshold, retrieval cache hit rate, and retrieval time of the repair impact feature retrieval cache at various levels, and to construct the final repair impact feature retrieval cache hit rate time mapping model. The current retrieval cache hit rate time-consuming mapping module is used to input the cache capacity data at each level corresponding to the current repair impact feature retrieval process and the semantic similarity threshold into the mapping model in the repair impact feature retrieval cache hit rate time-consuming mapping model construction module for mapping. The current retrieval semantic similarity threshold adjustment module is used to compare the mapping result in the current retrieval cache hit rate time mapping module with the corresponding preset threshold. If the requirements are met, no adjustment is needed; otherwise, the current semantic similarity threshold is repeatedly adjusted based on the mapping model in the repair impact feature retrieval cache hit rate time mapping model construction module, and the adjusted semantic similarity threshold is mapped again. If the mapping result meets the requirements within a limited number of times, the adjustment is completed; otherwise, the current retrieval cache capacity determination and adjustment module is executed. The current retrieval cache capacity determination and adjustment module is used to repeatedly adjust the retrieval cache capacity. It combines the adjusted semantic similarity threshold in the current retrieval semantic similarity threshold adjustment module with the mapping model in the feature retrieval cache hit rate time-consuming mapping model construction module to repair the mapping until the mapping result meets the requirements. Based on the finally determined cache capacity and similarity threshold, the retrieval operation is performed. After sorting the retrieval results, the iron-modified biochar is prepared and applied, and environmental risk and remediation benefit assessment is carried out.

[0022] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0023] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present invention.

Claims

1. A method for assessing the environmental risks and remediation benefits of iron-modified biochar, characterized in that, Includes the following steps: S1. Select several groups of historical iron-modified biochar remediation site remediation process effects and influence characteristics, and label and rank them. S2. Based on the labeled and sorted data in S1, construct the final sorted feature mapping model to repair the impact. S3. Obtain historical data on the cache capacity, semantic similarity threshold, cache hit rate, and retrieval time of several sets of environmental remediation impact feature retrieval data at various levels, and construct a final mapping model of cache hit rate and retrieval time for remediation impact feature retrieval. S4. Input the cache capacity data at each level corresponding to the current repair impact feature retrieval process and the semantic similarity threshold into the S3 mapping model for mapping; S5. Compare the mapping result in S4 with the corresponding preset threshold. If the requirements are met, no adjustment is needed. Otherwise, adjust the current semantic similarity threshold repeatedly based on the mapping model in S3, and map the adjusted semantic similarity threshold again. If the mapping result meets the requirements within the limited number of times, the adjustment is complete. Otherwise, execute S6; S6. Repeatedly adjust the retrieval cache capacity, and combine it with the semantic similarity threshold adjusted in S5 to input into the mapping model in S3 for mapping until the mapping result meets the requirements; perform the retrieval operation based on the finally determined cache capacity and similarity threshold, sort the retrieval results through the mapping model in S2, and then prepare and apply iron-modified biochar, and conduct environmental risk and remediation benefit assessment.

2. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 1, characterized in that, S1 includes the following steps: S11. Define several characteristic types that affect the environmental remediation effect of iron-modified biochar on contaminated sites, and obtain a set of biochar remediation impact characteristic types; then define several characteristic indicators that reflect the environmental remediation effect of iron-modified biochar, and obtain a set of biochar remediation effect characteristic types. S12. Based on the set of contaminated site feature types, the set of biochar remediation impact features, and the set of biochar remediation effect features, collect biochar remediation impact feature data and biochar remediation effect feature data corresponding to several historical instances of using iron-modified biochar for environmental remediation of contaminated sites, and obtain historical biochar remediation impact feature dataset and historical biochar remediation effect feature dataset. S13. Using the historical biochar remediation effect feature dataset, the remediation impact feature data of each group in the historical biochar remediation impact feature dataset are labeled to obtain the historical labeled biochar remediation impact feature dataset; the historical labeled biochar remediation impact feature dataset is randomly grouped to obtain the historical labeled remediation impact feature data set; Based on the historical biochar remediation effect feature dataset, and by manually sorting the data in each feature data group of the historical labeled remediation impact feature dataset according to the effect quality and removing the labeled data, the historical sorted remediation impact feature dataset is obtained.

3. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 2, characterized in that, S2 includes the following steps: S21. Based on the historical post-annotation repair impact feature data set and the historical post-sorting repair impact feature data set, construct a mapping model with the input of the post-annotation repair impact feature data set and the output of the post-sorting repair impact feature data set, and obtain the final post-sorting repair impact feature mapping model.

4. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 1, characterized in that, S3 includes the following steps: S31. During the process of searching for the impact features of iron-modified biochar remediation of contaminated sites, a multi-level retrieval system cache is set to obtain a remediation impact feature retrieval cache set; in conjunction with the remediation impact feature retrieval cache set, the total capacity data, retrieval semantic similarity threshold, retrieval cache hit rate data, and retrieval time data of the corresponding levels of contaminated site remediation impact feature retrieval cache are obtained in several historical processes of searching for the impact features of iron-modified biochar remediation of contaminated sites. Based on the collected data, a mapping model is constructed between retrieval cache capacity data, retrieval semantic similarity threshold and retrieval cache hit rate data, and retrieval time data, to obtain the final mapping model of retrieval cache hit rate and time consumption for repairing the impact of features.

5. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 4, characterized in that, S4 includes the following steps: S41. Label the data in the current iron-modified biochar environmental remediation impact feature knowledge base with remediation effect feature data; after the labeling is completed, obtain the total capacity data of the environmental remediation impact feature retrieval cache at each level during the current retrieval process and the corresponding preset retrieval semantic similarity threshold according to the remediation impact feature retrieval cache set, and obtain the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold. S42. Input the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model to obtain the current retrieval cache hit rate dataset and the current retrieval time data.

6. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 5, characterized in that, S5 includes the following steps: S51. Based on the characteristics of the impact of repair, retrieve the cache set and the current retrieval requirements, set the hit rate threshold of each level of cache in the current retrieval process and the current retrieval time threshold to obtain the current retrieval cache hit rate threshold set. S52. Set a first adjustment number threshold; if there is a cache hit rate data in the current retrieval cache hit rate dataset that is less than or equal to the hit rate threshold corresponding to the current retrieval cache hit rate threshold set, or if the current retrieval time data is greater than or equal to the current retrieval time threshold, the current retrieval semantic similarity threshold is repeatedly adjusted to obtain the current adjusted retrieval semantic similarity threshold. Otherwise, no adjustment is needed, and the current retrieval cache capacity dataset and the current retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold.

7. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 6, characterized in that, The S5 also includes: S53. The current adjusted retrieval semantic similarity threshold is combined with the current retrieval cache capacity dataset and input into the final repair feature retrieval cache hit rate time mapping model to obtain the current adjusted retrieval cache hit rate dataset and the current adjusted retrieval time data. If the number of repetitions in S52 is less than or equal to the first adjustment number threshold, and there is a cache hit rate data in the current adjusted retrieval cache hit rate dataset that is greater than the hit rate threshold corresponding to the current retrieval cache hit rate threshold set, and the current adjusted retrieval time data is less than the current retrieval time threshold, the adjustment is complete, and the current retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; otherwise, proceed to S6.

8. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 7, characterized in that, S6 includes the following steps: S61. Repeatedly adjust the current retrieval cache capacity dataset to obtain the current adjusted retrieval cache capacity dataset; input the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold into the final repair feature retrieval cache hit rate time mapping model for mapping to obtain the current second-adjusted retrieval cache hit rate dataset and the current second-adjusted retrieval time data. The adjustment is complete when there is a cache hit rate in the current second-adjusted retrieval cache hit rate dataset that is greater than the hit rate threshold corresponding to the current retrieval cache hit rate threshold set and the current second-adjusted retrieval time data is less than the current retrieval time threshold; the current adjusted retrieval cache capacity dataset and the current adjusted retrieval semantic similarity threshold are respectively used as the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold. S62. Adjust the cache capacity and semantic similarity threshold at each level during the current retrieval of the knowledge base data on the environmental remediation impact of iron-modified biochar based on the current final retrieval cache capacity dataset and the current final retrieval semantic similarity threshold; perform the retrieval operation after the adjustment is completed.

9. The method for assessing the environmental risk and remediation benefits of iron-modified biochar according to claim 8, characterized in that, S6 further includes: After the retrieval in S63 and S62 is completed, the current environmental remediation impact feature retrieval dataset is obtained; the current environmental remediation impact feature retrieval dataset is input into the final sorted remediation impact feature mapping model for mapping, and the current sorted environmental remediation impact feature retrieval dataset is obtained. Iron-modified biochar was prepared from the top-ranked data in the current sorted environmental remediation impact feature retrieval dataset. After preparation, it was applied to the corresponding contaminated site. After application, the environmental risk and remediation effect of the contaminated site were periodically assessed.

10. A system for implementing the method for assessing the environmental risks and remediation benefits of iron-modified biochar as described in any one of claims 1-9.

Images