Method for regulating and controlling optimization of dehalogenation carbon source material by utilizing machine learning to improve dehalogenation efficiency
Through machine learning, the preparation of carbon source materials is optimized, and the problem of mismatch between the added carbon source materials and the functional bacterial flora is solved, the dehalogenation efficiency and removal rate are improved, the cost is reduced, and the rapid and efficient removal of chlorinated olefins is achieved.
Patent Information
- Application Number
- CN202410973183.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-11-12
AI Technical Summary
The existing added carbon source materials do not match the physiological characteristics of the dechlorination functional flora, resulting in low dehalogenation efficiency and high economic cost, and limited understanding of the microbial reduction and dechlorination mechanism.
Using machine learning to regulate dehalogenated carbon source materials, identify the pores of key control factors acetic acid and carbon source materials through random forest algorithms, optimize the ratio of iron powder and humic acid, and prepare carbon source materials with small pores and high specific surface area, combined with agricultural waste and crosslinking agent solution to form a carbon source material suitable for functional microorganisms.
It significantly improves the dehalogenation efficiency, avoids secondary pollution, reduces economic costs, and maintains the microbial activity and the stability of carbon source materials, achieving rapid and efficient removal of chlorinated olefin pollutants.
Smart Images

Figure CN118942585A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of pollutant treatment, and specifically relates to a method for utilizing machine learning to regulate the optimization of dehalogenation carbon source materials to improve dehalogenation efficiency. Background Art
[0002] Bioremediation-stimulated degradation of halogenated hydrocarbons has been studied in the United States for nearly 40 years. Organohalogen compounds are widely distributed heterogeneous pollutants in the environment. Among these pollutants, chlorinated organics, including chlorinated hydrocarbons (CEs), chlorinated alkanes (CAs), and polychlorinated biphenyls (PCBs), are typical environmental pollutants with environmental sustainability, bioaccumulation, and toxicity. Microbial-mediated remediation of halogenated hydrocarbons has become a promising remediation method due to its strong adaptability and wide range of remediation. Compared with the physiological sensitivity and metabolic complexity of obligate organohalide respiring bacteria (OHRB), iron-reducing bacteria (S. oneidensis MR-1) are known for their diverse respiratory abilities and have been widely used in bioremediation efforts. This functional microorganism has been shown to be capable of reductive dechlorination via extracellular electron transfer (EET). In addition, it can utilize organic matter such as humic acid (HA) as an extracellular electron transfer mediator to promote the removal of halogenated pollutants through iron-mediated redox reactions. Therefore, it plays a vital role in the remediation of wastewater containing iron and humic substances. However, the carbohydrate metabolism capacity of this type of microorganism is very limited. A large number of studies have shown that it can only use simple organic matter as electron donors, such as acetic acid and lactic acid, for growth metabolism, which limits its application in in situ bioremediation.
[0003] Soluble electron donors including organic acids and alcohols have been shown to be effective electron donors for the removal of halogenated hydrocarbon contaminants from contaminated sites. However, their introduction may lead to pore blockage in aquifers due to excessive cell growth, which may cause secondary contamination. As an alternative, cellulose-rich agricultural waste carbon sources offer a sustainable and environmentally friendly option for site bioremediation technology due to their wide availability and good biocompatibility. However, these carbon sources have limited utilization in environmental bioremediation efforts due to their slow electron release capacity. Iron-humic acid (Fe HA) complexes are ubiquitous in aquatic environments and have been shown to promote the hydrolysis of complex compounds by forming various iron oxides. Therefore, we proposed the use of Fe-HA complexes to couple complex carbon sources with the metabolism of iron-reducing bacteria, aiming to improve dehalogenation efficiency while reducing economic costs and achieve sustainable bioremediation. Nevertheless, this approach is affected by many complex factors, and the understanding of the mechanism of microbial reductive dechlorination is still limited.
[0004] Machine learning has become a powerful tool for analyzing big data and shows great potential for enhancing bioremediation management. It has been applied to predict the vulnerability of groundwater contaminants, assess microbial diversity, and identify molecular features that contribute to material properties, such as molecular formula or elemental composition. In addition, multi-omics analysis provides more comprehensive physiological and functional information of microorganisms, which helps to explore their metabolic functions. Therefore, based on machine learning and transcriptomics methods, it is of great significance to analyze the functional activities of functional microorganisms that degrade chlorinated hydrocarbon contaminated sites to explore the mechanism of biostimulation. However, there is still a gap in the combination of multi-omics analysis to guide the rational design of carbon sources to optimize dehalogenation strategies in groundwater bioremediation applications. There is an urgent need to design dehalogenation materials based on microbial physiological responses to achieve high dehalogenation efficiency efficiently and accurately. Summary of the invention
[0005] The purpose of the present invention is to solve the problem that the existing external carbon source materials are not compatible with the physiological characteristics of the dechlorination functional bacteria, and to provide a method for optimizing the dehalogenation carbon source materials by using machine learning to improve the dehalogenation efficiency.
[0006] The present invention discloses a method for optimizing dehalogenation carbon source materials by using machine learning to improve dehalogenation efficiency. The preparation method is carried out according to the following steps:
[0007] A random forest was selected as a machine learning model, and the dehalogenated carbon source material was subjected to biostimulation to obtain significantly different functional genes, carbon source material synthesis parameters and water quality conditions as input, and dehalogenation efficiency as output; the data set was randomly divided into 80% and 20% using a random forest algorithm for training and testing, respectively; the key control factors under biostimulation conditions were identified using a random forest algorithm, and verification experiments were designed based on the identified key control factors to optimize the ratio of iron powder and humic acid in the carbon source material synthesis parameters, thereby completing the optimization of dehalogenated carbon source materials regulated by machine learning; wherein the key control factors were acetic acid and the pores of the carbon source material.
[0008] Furthermore, the carbon source material has a pore size of ≤100 nm and an acetic acid content of ≥3000 mg / L.
[0009] The carbon source material has a pore size of 50-100 nm and an acetic acid content of 2500-3200 mg / L.
[0010] Furthermore, the synthesis parameters of the carbon source material are the carbon percentage, oxygen percentage, ferric iron percentage, ferrous iron percentage, carbon source material porosity and specific surface area.
[0011] Furthermore, the water quality conditions are COD, DOC, pH, redox potential, electron transfer capacity and volatile acid types.
[0012] Furthermore, the dehalogenated carbon source material is prepared as follows:
[0013] 1. Preparation of cross-linking agent solution:
[0014] Dissolve polyvinyl alcohol and sodium alginate in distilled water by heating, heat in a water bath at 95-100° C. and stir for 6-8 hours until completely dissolved, and cool to room temperature to obtain a crosslinking agent solution;
[0015] 2. Preparation of carbon source materials suitable for reduction and dehalogenation:
[0016] Add diatomaceous earth, agricultural waste carbon source material powder, iron powder, and finally humic acid powder to the crosslinking agent solution, mix and extrude into spherical particles, and then add CaCl 2 The method comprises the following steps: cross-linking in a saturated boric acid solution for 24 to 48 hours, drying and sterilizing to obtain a carbon source material suitable for reduction dehalogenation; wherein the mass content of the agricultural waste carbon source material in the carbon source material is 5% to 10%; and the mass volume ratio of diatomaceous earth to the cross-linking agent solution is (1-4) g: (2-8) mL.
[0017] Furthermore, the agricultural waste is one or more of rice husks, corn cobs, straws, peanut shells, and wood chips.
[0018] Furthermore, the agricultural waste carbon source material is obtained by pre-treating the agricultural waste by soaking it in a NaOH solution for 4 to 8 hours, drying it, and crushing it into powder.
[0019] Furthermore, the mass volume ratio of polyvinyl alcohol, sodium alginate and deionized water is (8-16) g: (1-2) g: (300-600) mL.
[0020] Furthermore, the mass volume ratio of iron powder to deionized water is (10-20) g: (1000-2000) mL, wherein the amount of iron powder added is 10-15 g / L.
[0021] Furthermore, the mass volume ratio of humic acid powder to deionized water is (18-36) g: (1000-2000) mL, wherein the dosage of humic acid powder is 20-40 g / L.
[0022] According to the identified key control factor acetic acid, the mass volume ratio of iron powder and humic acid powder is adjusted to (10-20) g: (18-36) g, the iron powder is used to stimulate the hydrolysis of macromolecular carbon sources, and the humic acid is used to increase the diffusion rate of small molecular substance acetic acid, wherein the addition amount of iron powder is 10-15 g / L, and the addition amount of humic acid is 20-40 g / L.
[0023] Random forest was selected as the machine learning model. The 831 significantly different functional genes obtained by biostimulation of the above carbon source materials, carbon source material synthesis parameters (C%, O%, Fe%, Fe(III)%, Fe(II)%, carbon source material pores, specific surface area) and water quality conditions (COD, DOC, pH, redox potential, electron transfer capacity, volatile acid types (ethanol, acetic acid, propionic acid, pyruvic acid and lactic acid) were used as input, and dehalogenation efficiency was used as output. The data set was randomly divided into 80% and 20% for training and testing, respectively, using the random forest algorithm. The key control factors (acetic acid and carbon source material pores) under biostimulation conditions were identified using the random forest algorithm. According to the identified key control factors, a verification experiment was designed to optimize the ratio of iron powder and humic acid in the carbon source material, and a dehalogenated carbon source material with a carbon source material pore of 50-100nm and capable of producing a large amount of acetic acid was prepared. The carbon source material achieved the purpose of matching the physiological characteristics of functional microorganisms.
[0024] The present invention is based on agricultural waste carbon source, and the carbon source material matched with the physiological characteristics of functional microorganisms prepared by the method of machine learning not only retains the advantages of agricultural waste carbon source such as high efficiency, strong adsorption, good carbon release effect, etc., which is suitable for purifying water quality, but also has the advantages of high efficiency in adsorbing chlorinated olefin pollutants, fast rate and strong stability. Under the preparation conditions, agricultural waste is used as the basic raw material of the dehalogenated carbon source, iron powder and humic acid are added, and a 60-day high-efficiency degradation experiment of chlorinated olefins is carried out. The experimental conditions are set to rice husk + iron powder + humic acid, corn cob + iron powder + humic acid, straw + iron powder + humic acid, peanut shell + iron powder + humic acid, sawdust + iron powder + humic acid. The results showed that the carbon release effects of the five agricultural waste carbon source materials were relatively stable, and the maximum COD release in 60 days could reach ~85.94-87.66 mg / d (corn cob + iron powder + humic acid and sawdust + iron powder + humic acid). The increase in COD concentration indicated that the carbon source material could be used as a carbon substrate to maintain microbial activity, so that chlorinated olefin pollutants could be quickly and efficiently degraded while avoiding secondary pollution. And the pH could be kept stable within 60 days, indicating that the carbon source material would not cause groundwater acidification and had a good buffering effect. In addition, the removal rate of chlorinated olefins continued to increase with the extension of reaction time. The removal rates of chlorinated olefins under the conditions of sawdust + iron powder + humic acid were 1.15 times, 1.34 times, 1.16 times and 1.26 times of those under the conditions of rice husk + iron powder + humic acid, corn cob + iron powder + humic acid, straw + iron powder + humic acid, and peanut shell + iron powder + humic acid, respectively. This shows that sawdust exhibits better bioremediation effects as a carbon source material. Therefore, the present invention then selected sawdust + iron powder + humic acid as electron donors for multi-omics experiments. It is worth noting that, unlike the biochar preparation method, the carbon source material is prepared by chemical crosslinking, which does not require high-temperature calcination, avoids the performance difference of biochar caused by the influence of pyrolysis temperature, significantly saves economic costs and reduces greenhouse gas emissions. And chemical crosslinking can improve the performance of the material, improve the strength and heat resistance of the carbon source material. In addition, iron powder is a strong reducing agent, so it has excellent adsorption performance and high reduction activity, and can remove chlorinated olefin pollutants by reduction under non-biological conditions; however, under biological stimulation conditions, iron powder will produce a large number of strong reducing free radicals to destroy the physiological structure of cells, cause the outflow of cell contents, and affect the physiological function of microorganisms. In addition, humic acid is a common natural compound containing a large number of quinones, hydroxyl groups, carboxyl groups and other functional groups, which can be used as an extracellular electron transfer shuttle to accelerate electron transfer; in addition, the solubilization of humic acid can increase the diffusion rate of water-soluble substances in the diffusion layer around bacterial cells, thereby promoting their absorption by bacteria. More importantly, humic acid can also affect the environment of iron by forming soluble complexes or aggregates on the mineral surface.Observations show that these iron-humic acid complexes affect the solubility, toxicity and fluidity of metals by forming various iron (hydroxide) oxides, including iron hydride, goethite and hematite. The addition of humic acid can alleviate the physiological inhibitory effect of iron powder on microorganisms to a certain extent, quickly provide electrons to accelerate the reduction of pollutants while reducing the toxic inhibition of chlorinated olefin pollutants, and significantly reduce the economic cost of carbon source materials. In addition, polyvinyl alcohol (PVA) and sodium alginate (SA) are used as cross-linking carriers in the preparation process of the carbon source material. They are colloidal carbohydrates with stable structures. Under the action of the cross-linking agent, the macromolecular chains are linked by chemical bonds to form a three-dimensional network structure of polymer polymers. This structure makes the material structure compact and not easy to hydrolyze. And this three-dimensional network structure can play a certain protective role on the functional bacteria in polluted water bodies. At the same time, the saturated boric acid solution contains 4% calcium chloride solution, Ca. + The ions are evenly distributed throughout the solution. Boric acid dehydrates with -OH on polyvinyl alcohol due to the lack of electrons to form a stable complex, and the prepared gel particles have a very uniform structure. In addition, diatomaceous earth is added during the preparation of the slow-release carbon source. Diatomaceous earth is formed by the accumulation of single-cell biological skeletons. It has a porous structure that provides space for strains in the water to attach, which can improve the adsorption performance of the material. In the case of dehydration, the structure is more compact, and its surface contains a large number of Si-OH silanol groups. During the ionization or hydrolysis process, hydroxyl radicals -OH are generated to form negative charges on the surface of diatomaceous earth. This increases the surface porosity and carbon distribution of the carbon source material after carbon is released during the carbon release process. The exposed -OH quickly decomposes the pollutants and forms a stable complex with the cross-linking agent boric acid B and -OH, forming a stable mutual protection between carbon and silicon on the surface of the material. This interaction dominates the carbon release stability and adsorption performance of the material, while maintaining the activity of microorganisms.
[0025] Regarding the machine learning aspect, specifically, the raw data selected as input features in the present invention are collected from the biostimulation reduction dechlorination experiment conducted by the present invention. The data set contains 831 significantly expressed differential genes, carbon source material synthesis parameters (C%, O%, Fe%, Fe(III)%, Fe(II)%, carbon source material pores, specific surface area) and water quality conditions (COD, DOC, pH, redox potential, electron transfer capacity, volatile acid species), and dechlorination efficiency is the only output variable. Random forests were selected to develop the classification model, which was mainly developed and implemented by the R software package (4.0.3) "randomForest". The data set in this study was randomly divided into 80% and 20% for training and testing, respectively. Before establishing each model, the optimal value was determined by 10-fold cross validation to improve the prediction efficiency. The feature importance of the optimized model was calculated and visualized by the "caret" and "ggplot2" packages. The importance of the input parameters was ranked by machine learning. According to the machine learning model, we obtained that the key control factors affecting the dechlorination effect were related to the carbon metabolic pathway of functional microorganisms. The three most critical metabolic features were extracted, namely: amino acid biosynthesis, carbon metabolism and fatty acid metabolism. The metabolic pathways were correlated with the carbon source synthesis parameters, and the results showed that COD had a significant positive effect on the carbon metabolic pathway, indicating that the excellent dechlorination performance of the system under the stimulation of carbon source materials originated from the effective absorption of carbon sources by the strain. In addition, small molecule acetic acid and dechlorination efficiency are closely related to carbon metabolism functional genes and carbon source synthesis parameters. The carbon source material pores of the material showed a significant positive correlation with amino acid biosynthesis and fatty acid metabolism, and the results showed that smaller carbon source material pores can significantly improve the dechlorination effect. Therefore, these findings show that the enhancement of the carbon metabolism ability of the strain can be attributed to Fe (III) as an electron acceptor, which accelerates the circulation of iron oxides, thereby enhancing the process of extracellular electron transfer of the strain. The enhancement of carbon metabolic energy also promotes the physiological activity of the strain, thereby accelerating the ability of bacteria to use chlorinated pollutants as electron acceptors, and promoting the efficient removal of pollutants. Based on these findings, the present invention designs a carbon source material with smaller carbon source material pores (50-100nm) and can produce acetic acid in large quantities for verification experiments.
[0026] Based on the results of the verification test, it is confirmed that regulating the composition ratio of iron powder and volatile acid in the carbon source material based on agricultural waste can control the pore size of the carbon source material prepared, and regulate the production of small molecule volatile acid acetic acid. As an electron donor preferred by functional strains, the production of acetic acid can significantly promote the dehalogenation effect of the carbon source material system improved based on machine learning. In addition, compared with the previous dehalogenation carbon source material that only focuses on improving the physiological metabolic activity of microorganisms, the carbon source material prepared by the present invention improves the physiological metabolism of microorganisms while the smaller carbon source material pores enhance the dispersibility of the carbon source material, and the large number of pores enhance the physical adsorption of pollutants, which increases the dehalogenation effect by 2.5 times, and has obvious cost advantages and application prospects. Therefore, the present invention is based on a method for optimizing the dehalogenation carbon source material by using machine learning to continuously, stably and efficiently provide energy and electrons to iron-reducing bacteria for a long time by stably producing acetic acid, thereby enhancing the physiological activity of iron-reducing bacteria and the ability of extracellular electron transfer, thereby improving the removal effect of chlorinated olefins; at the same time, the dehalogenation carbon source prepared by the present invention using agricultural waste as raw materials has a rough flaky loading layer, which increases the adsorption sites for iron-reducing bacteria to contact chlorinated olefins; in the absence of humic acid, iron powder undergoes an oxidation-reduction reaction during the reduction and dechlorination process to generate magnetite Fe with larger carbon source material pores. 2 O 3 , and under the action of humic acid oxygen-containing functional groups, the sawdust-iron-humic acid system produces goethite FeOOH with smaller carbon source material pores and stronger adsorption capacity, which promotes the continuous reduction of Fe(III). Therefore, the present invention achieves the purpose of regulating the type of iron oxide in the system by regulating the ratio of iron powder and humic acid, thereby affecting the pore size of the carbon source material and the production of acetic acid. The carbon source material prepared by the present invention has a larger specific surface area and a smaller carbon source material pore (≤100nm), and a large number of tiny pores enhance the adsorption capacity of pollutants. Together with acetic acid (≥3000mg / L) to accelerate the physiological metabolism of microorganisms, it significantly promotes the continuous and efficient degradation of extracellular pollutants. Compared with the existing dehalogenated carbon source materials, the carbon source materials prepared by the present invention using machine learning methods are more compatible with the physiological functions of microorganisms, and the dehalogenation effect is more sustained and stable, so a faster and more efficient pollutant removal effect can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 This is a graph showing the removal effect of chlorinated olefins from five agricultural waste carbon sources;
[0028] Figure 2 This is a graph showing the carbon release patterns of five agricultural waste carbon source materials;
[0029] Figure 3 The volatile acid distribution diagram of five agricultural waste carbon source materials;
[0030] Figure 4 This is the pH variation diagram of five agricultural waste carbon sources;
[0031] Figure 5 Identify key metabolic pathway graphs for the random forest algorithm;
[0032] Figure 6 It is the correlation diagram between functional metabolic pathways and carbon source synthesis parameters;
[0033] Figure 7 To verify the experimental regulation of acetic acid generation diagram and vinyl chloride degradation effect diagram;
[0034] Figure 8 The specific surface area and pore size distribution of the constructed composite carbon source material;
[0035] Fig. 9 XRD pattern for identification of iron oxide species;
[0036] Fig.10 This is a diagram to verify the degradation effect of vinyl chloride in the experiment. DETAILED DESCRIPTION
[0037] In order to make the objectives, technical solutions and advantages of the embodiments of the present invention more clearly understood, the spirit of the contents disclosed by the present invention will be described in detail below. After understanding the embodiments of the contents of the present invention, any technician in the relevant technical field can change and modify them according to the techniques taught by the contents of the present invention without departing from the spirit and scope of the contents of the present invention.
[0038] The exemplary embodiments of the present invention and the description thereof are used to explain the present invention but are not intended to limit the present invention.
[0039] The beneficial effects of the present invention are verified by the following experiments:
[0040] Embodiment 1,
[0041] A method of optimizing dehalogenated carbon source materials by using machine learning in this embodiment:
[0042] Agricultural waste materials: rice husks, corn cobs, straw, peanut shells, sawdust were washed, then soaked in NaOH solution for 4 hours for pretreatment, placed in an 80℃ oven for drying for 24 hours, taken out and cooled, crushed into powder with a particle size of 200-400 mesh using a Yongli brand crusher for standby use, and agricultural waste powder was obtained; 72g of polyvinyl alcohol and 9g of sodium alginate were weighed and heated to dissolve in distilled water, heated and stirred in a water bath at 95-100℃ for 4-8 hours until completely dissolved, cooled at room temperature, and a crosslinker solution (polyvinyl alcohol PVA-sodium alginate SA carrier skeleton) was obtained; 2.4g of agricultural waste powder was weighed and placed in a 100mL beaker, 40mL of polyvinyl alcohol PVA-sodium alginate SA carrier skeleton, 20g of diatomaceous earth, 0.6g of nano zero-valent iron, and 0.72g of humic acid were added, and mixed and stirred. A uniformly mixed slow-release carbon source mixed system containing natural polymer sawdust was prepared. Then, prepare 1000 mL of 4% CaCl 2 The saturated boric acid solution was cooled and dispensed into 100 mL conical flasks, squeezed into spherical particles with a particle size of about 0.7 mm, and placed in 4-8% CaCl 2 Keep the spherical cross-linking in saturated boric acid solution for 24 hours, take out and dry in an oven at 80°C for 24 hours, and sterilize with UV for 72-96 hours for use;
[0043] In this embodiment, five carbon source materials for biostimulation were prepared using rice husk + iron powder + humic acid, corn cob + iron powder + humic acid, straw + iron powder + humic acid, peanut shell + iron powder + humic acid, and sawdust + iron powder + humic acid. The obtained carbon source materials were named RH-Fe-HA, CC-Fe-HA, WS-Fe-HA, PS-Fe-HA, and SD-Fe-HA. The removal effects of chloroolefins from the five agricultural waste carbon sources in 60 days were as follows: Figure 1 As shown by Figure 1 It can be seen that on the 60th day, all carbon source materials (RH-Fe-HA, CC-Fe-HA, WS-Fe-HA, PS-Fe-HA, SD-Fe-HA) can maintain a good pollutant removal effect. At the same time, the removal rate of SD-Fe-HA for chlorinated olefins can even reach 99.8%, which is 1.15 times, 1.34 times, 1.16 times and 1.26 times that of RH-Fe-HA, CC-Fe-HA, WS-Fe-HA and PS-Fe-HA respectively, indicating that the carbon source material with sawdust (SD) as raw material can significantly improve the conversion efficiency of chlorinated olefin pollutants and enhance the application effect of carbon source materials after adding iron powder and humic acid.
[0044] The carbon release law of the slow-release carbon source material prepared in this example is as follows: Figure 2 As shown, from Figure 2It can be seen that as the reaction time increases, the carbon release effect of all carbon source materials increases gradually. On the 60th day, the COD release can reach a maximum of ~85.94-87.66 mg / d (corn cob + iron powder + humic acid and sawdust + iron powder + humic acid). This shows that the prepared carbon source materials have good carbon release ability. Figure 3 The distribution of volatile acids of five agricultural waste carbon source materials. At the 60th day, acetic acid and ethanol were the main fermentation products. Compared with other materials, corn cob + iron powder + humic acid (CC-Fe-HA) showed 2.3-11.8 times higher ethanol yield. Due to the acidification effect caused by the rapid release of carbon in the corn cob + iron powder + humic acid system, it cannot be considered as the best carbon source material ( Figure 4 ). Figure 5 Identify key metabolic pathways for the random forest algorithm, from Figure 5 As can be seen from the results, the four most important KEGG pathways affecting dechlorination (p < 0.001) include global and overview maps, carbohydrate metabolism, amino acid metabolism, and translation. It is worth noting that the global and overview map pathways are considered to be crucial and are mainly related to carbon metabolism, amino acid biosynthesis, and fatty acid metabolism. In particular, fatty acid metabolism is very important for predicting the potential for reductive dechlorination. Small molecular fatty acids can catalyze the key steps in the dechlorination process. In addition, during reductive dechlorination, small molecular volatile acids can be converted into acetic acid and H 2 , acting as electron donors for anaerobic reductive dechlorination and promoting the degradation of chlorinated olefins. Therefore, when designing composites, it is crucial to improve the functional categories related to carbon metabolism, which will affect the subsequent energy production and substrate transport.
[0045] 2. Activation of iron-reducing bacteria Shewanella oneidensis MR-1 using the above agricultural waste carbon source (ATCC BAA-1096)
[0046] Pre-culture of iron-reducing bacteria Shewanella oneidensis MR-1: prepare solid LB culture medium. Use a pH meter to adjust the pH of the prepared culture medium to 7.2-7.4, sterilize with high-pressure steam at 121°C for 20-30 minutes, cool to 60-65°C, pour into a plate for use, and place in a constant temperature incubator at 30-35°C for 32-48 hours. Then inoculate in liquid LB culture medium for expanded culture, where the activation culture medium formula is: 10g / L tryptone, 5g / L yeast extract and 10g / L sodium chloride, and adjust the pH value of the solution to 7.2-7.4 with 1% NaOH. Resuspend the bacterial liquid of iron-reducing bacteria Shewanella oneidensisMR-1 that has been cultured for 36-48 hours. The specific steps are to take 10mL of liquid, centrifuge at 5500*g for 5 minutes, and wash three times with sterile water. Cultivate for 48 hours to make OD 600 The inorganic salt medium contains 50mM BES-Na sodium salt buffer (1.175g), 10mM PBS buffer solution (Na 2 HPO 4 12H 2 O (3.58 g / L), NaH 2 PO 4 · 2 H 2 O(1.78 g / L)), NH 4 Cl (0.31 g / L), KCl (0.13 g / L)), 12.5 mL of metal elements, and sodium lactate (2.24 g / L) were used as electron donors. The carbon source of the culture medium was replaced with the prepared agricultural waste carbon source material, and 50 μmol / L of the target pollutant chlorinated olefin (tetrachloroethylene) was added as an electron acceptor. In addition, the culture medium was supplemented with amino acid and trace mineral stock solutions. The pH value of the culture medium was adjusted to 7.2-7.4 with NaOH. Subsequently, 20% and OD 600 2.0-2.5 of Shewanella oneidensis MR-1 bacteria were inoculated into 150 mL brown anaerobic bottles and allowed to stand for 60 days. Water samples were taken on the 10th, 20th, 30th, 40th, 50th and 60th days after the addition of the carbon source to analyze pH, chemical oxygen demand (COD) and gas chromatography (GC).
[0047] The treatment conditions in Example 1 are as follows: the carbon source dosage is 1-1.5 g, added to 150 mL of laboratory water, the pH value is adjusted to 7.2-7.4, and placed in a brown, sealed, oxygen-free anaerobic bottle. The biostimulation experiment was carried out using the carbon source material (SD-Fe-HA) with the best dechlorination effect. During the experiment, the release of chemical oxygen demand tended to be stable, which shows that the carbon release tended to be balanced, indicating that sawdust (SD) as the basic material of the carbon source can release electron donors and carbon sources to maintain the growth and energy metabolism of organisms for the degradation of pollutants. At this time, most of the organic carbon in the water comes from the macromolecular cellulose substances produced by the hydrolysis of sawdust materials. And COD continues to rise with time, which shows that the addition of this carbon source shows good carbon release performance. Figure 6 It is a correlation diagram between functional metabolic pathways and carbon source synthesis parameters. The results show that COD has a significant positive effect on carbon metabolic pathways, indicating that under the stimulation of carbon source materials, the excellent dechlorination performance of the system is derived from the effective absorption of carbon source by the strain. In addition, small molecule acetic acid and dechlorination efficiency are closely related to carbon metabolism functional genes and carbon source synthesis parameters. Fe (II) and Fe% show a significant positive correlation with amino acid biosynthesis and fatty acid metabolism. Therefore, these findings show that the enhancement of the carbon metabolism ability of the strain can be attributed to Fe (III) as an electron acceptor, thereby enhancing the process of extracellular electron transfer. The enhancement of carbon metabolism energy promotes the physiological activity of the strain, thereby accelerating the ability of bacteria to use chlorinated pollutants as electron acceptors, and promoting the efficient removal of pollutants. Based on these findings, the present embodiment designs a carbon source material with a smaller carbon source material pore (50-100nm) and a large amount of acetic acid to carry out a verification test.
[0048] Figure 7 In order to verify the experimental regulation of acetic acid generation and vinyl chloride degradation effect, Figure 7 It can be seen that compared with the humic acid adjustment group, the acetic acid content in the Fe adjustment group increased significantly with the increase of humic acid concentration. This finding emphasizes that Fe plays a key role in driving carbon source hydrolysis, and the oxygen-containing functional groups of humic acid can stimulate the hydrolysis of complex carbon sources to produce more small molecular acetic acid. Therefore, the iron powder and humic acid regulate the production of acetic acid through a strong interaction relationship. In addition, compared with the acetic acid yield in the experimental study before optimization (103.3-394.6 mg / L*d), the acetic acid yield obtained under the conditions obtained by the optimization model based on machine learning increased by about ~30% ( Figure 7 ). Figure 8 The specific surface area and pore size distribution of the constructed composite carbon source material. It can be seen from the figure that the carbon source material prepared by adjusting the proportion of humic acid has a larger specific surface area and a smaller carbon source material pore distribution (≤100nm). Under the condition of no addition of humic acid, the iron powder undergoes redox reaction during the reduction and dechlorination process to generate magnetite Fe with larger carbon source material pores. 2O 3 Under the action of the oxygen-containing functional groups of humic acid, the carbon source material produced by the sawdust-iron-humic acid system has smaller pores and stronger adsorption capacity, which promotes the continuous reduction of Fe(III) ( Fig. 9 The smaller pore distribution of the carbon source material improves the adsorption capacity of pollutants, and the larger specific surface area is the adsorption site for iron-reducing bacteria to contact chlorinated olefins, which significantly promotes the continuous and efficient degradation of extracellular pollutants and improves the dehalogenation effect by 2.5 times ( Fig.10 In addition, Fe conditioning and humic acid conditioning achieved 99.4% and 94.7% perchloroethylene removal, respectively, within 20 days ( Fig.10 ). This shows that the reduction dechlorination effect is indeed achieved by increasing the content of acetic acid in the system and the smaller carbon source material pores. These results are consistent with the findings of the transcriptome, confirming that acetic acid has a significant contribution to dechlorination. The experiment of this embodiment emphasizes the importance of carbohydrate metabolic pathways in promoting carbon source utilization and accelerating reduction dechlorination. When the ratio of sawdust-iron powder-humic acid is 2.4g:0.4g:1.44g, a carbon source material with smaller carbon source material pores (≤100nm) and capable of producing a large amount of acetic acid (≥3000mg / L) can be produced, successfully achieving the goal of accurately designing physiologically matched carbon source materials using the physiological characteristics of functional bacteria. Therefore, the present invention successfully constructs a dehalogenated carbon source material that matches the physiological characteristics of iron-reducing bacteria using a machine learning method, providing a new idea for in-situ bioremediation of chlorinated olefin contaminated sites.
Claims
1. A method for optimizing dehalogenation carbon source materials by using machine learning to improve dehalogenation efficiency, characterized in that The preparation method is carried out according to the following steps: A random forest was selected as a machine learning model, and the dehalogenated carbon source material was subjected to biostimulation to obtain significantly different functional genes, carbon source material synthesis parameters and water quality conditions as input, and dehalogenation efficiency as output; the data set was randomly divided into 80% and 20% using a random forest algorithm for training and testing, respectively; the key control factors under biostimulation conditions were identified using a random forest algorithm, and verification experiments were designed based on the identified key control factors to optimize the ratio of iron powder and humic acid in the carbon source material synthesis parameters, thereby completing the optimization of dehalogenated carbon source materials regulated by machine learning; wherein the key control factors were acetic acid and the pores of the carbon source material.
2. The method of claim 1, wherein the method comprises: The carbon source material has a pore size of ≤100 nm and an acetic acid content of ≥3000 mg / L.
3. The method of claim 1, wherein the method comprises: The synthesis parameters of the carbon source material are the carbon percentage, oxygen percentage, ferric iron percentage, divalent iron percentage, pores and specific surface area of the carbon source material.
4. The method of claim 1, wherein the method comprises: The water quality conditions are COD, DOC, pH, redox potential, electron transfer capacity and volatile acid types.
5. The method of claim 1, wherein the method comprises: The dehalogenated carbon source material is prepared as follows:
1. Preparation of crosslinker solution: Dissolve polyvinyl alcohol and sodium alginate in distilled water by heating, heat in a water bath at 95-100° C. and stir for 6-8 hours until completely dissolved, and cool to room temperature to obtain a crosslinking agent solution; 2. Preparation of carbon source materials suitable for reduction and dehalogenation: Add diatomaceous earth, agricultural waste carbon source material powder, iron powder and finally humic acid powder to the crosslinking agent solution, extrude the mixture into spherical particles, put them into a CaCl2 saturated boric acid solution, crosslink for 24 to 48 hours, dry and sterilize to obtain a carbon source material suitable for reduction dehalogenation; wherein the mass content of the agricultural waste carbon source material in the carbon source material is 5% to 10%; and the mass volume ratio of diatomaceous earth to the crosslinking agent solution is (1-4) g: (2-8) mL.
6. The method of claim 5, wherein the method comprises: The agricultural waste is one or more of rice husks, corn cobs, straws, peanut shells and wood chips.
7. A method for improving dehalogenation efficiency by using machine learning to control the optimization of dehalogenation carbon source materials according to claim 5 or 6, characterized in that The agricultural waste carbon source material is obtained by pre-treating the agricultural waste by soaking it in a NaOH solution for 4 to 8 hours, drying it, and crushing it into powder.
8. The method of claim 5, wherein the method comprises: The mass volume ratio of polyvinyl alcohol, sodium alginate and deionized water is (8-16) g: (1-2) g: (300-600) mL.
9. The method of claim 5, wherein the method comprises: The mass volume ratio of iron powder and deionized water is (10-20) g: (1000-2000) mL, wherein the amount of iron powder added is 10-15 g / L.
10. The method of claim 5, wherein the method comprises: The mass volume ratio of humic acid powder and deionized water is (18-36) g: (1000-2000) mL, wherein the dosage of humic acid powder is 20-40 g / L.
