A method and system for predicting porosity based on iron ore fines and mineral composition information

Abstract

The application discloses a kind of porosity prediction method and system based on iron ore powder and mineral composition information, belong to sinter production field.The application first determines the chemical composition and mineral composition of single iron ore powder, and combines its standard roasting experimental data, establishes single ore porosity prediction model;Further, according to the actual ore proportioning ratio, the theoretical porosity of mixed ore is calculated;Finally, the final model for predicting the actual sinter porosity is constructed by comprehensively considering the theoretical porosity of mixed ore and the sintering process parameters.The method solves the technical problem that the traditional method cannot accurately and quantitatively predict the sinter porosity before production, realizes the prospective optimization and adjustment of ore proportioning structure and process parameters based on the prediction results, so as to stabilize the sinter porosity in the best range, so as to improve the sinter quality.

Description

Technical Field

[0001] This invention relates to the field of sinter production technology, specifically to a porosity prediction method and system based on iron ore powder and mineral composition information. Background Technology

[0002] In steel enterprises, during the iron ore sintering process, various raw materials are mixed and granulated, then distributed to form a granular bed with suitable permeability. Under high temperature, the materials partially melt and bind together, ultimately producing a porous, blocky finished sinter. The pore structure formed inside the sinter directly affects its mechanical strength and reducibility, and ultimately the efficiency and energy consumption of blast furnace ironmaking. Therefore, obtaining and controlling a reasonable porosity is crucial for optimizing the entire sintering and blast furnace production process.

[0003] However, in actual production, the porosity of sinter is an indicator that is difficult to predict accurately before forming. It is determined by the chemical composition and mineral composition of the raw materials, as well as complex sintering process parameters. Traditional methods mainly rely on testing after sintering to obtain the results. This reactive approach leads to a lag in production adjustments, making it impossible to proactively and forward-lookingly control the porosity within the ideal range through ore blending and process optimization. This affects the stability of sinter quality and the operation of the blast furnace.

[0004] Chinese patent CN107341289A discloses a calculation method for describing the change in porosity of iron ore sintering bed. By simulating the change law of porosity in each zone during sintering, the sensitivity of influencing factors is analyzed to improve the accuracy of numerical calculation of sintering mass-heat process. However, this scheme is a calculation performed during sintering and cannot be used for pre-production guidance.

[0005] Chinese patent CN118097346A discloses a multimodal fusion method and system for online identification of blast furnace raw material particle size. This method uses advanced technologies such as 3D point cloud and deep learning to identify and segment sintered ore products online in order to achieve particle size statistics and morphology analysis. However, it is still a post-event detection method and cannot achieve pre-event prediction. Summary of the Invention

[0006] The purpose of this invention is to provide a porosity prediction method and system based on iron ore powder and mineral composition information. By measuring the chemical composition and mineral composition of a single type of iron ore powder and combining it with its standard roasting experimental data, a porosity prediction model for a single ore is established. Then, the theoretical porosity of the blended ore is calculated according to the actual blending ratio. Finally, by combining the theoretical porosity of the blended ore and the sintering process parameters, a final model for predicting the porosity of the actual sinter is constructed. This enables the forward-looking optimization and adjustment of the blending structure and process parameters based on the prediction results, thereby stabilizing the porosity of the sinter within the optimal range, improving the quality of the sinter, and solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A porosity prediction method based on iron ore powder and mineral composition information includes the following steps:

[0009] S1. Obtain the content of major chemical components and the mass fraction of major mineral composition of various single-variety iron ore powders. The major chemical components include at least TFe. CaO The main mineral composition includes at least magnetite, hematite, limonite, and quartz, and consists of MgO and FeO.

[0010] S2. Each type of iron ore powder was roasted in the laboratory under a fixed binary basicity R, and its porosity after roasting was measured.

[0011] S3. Using the main chemical components and the mass fraction of the main mineral composition of each single iron ore powder as independent variables and the corresponding roasted porosity as dependent variable, the partial least squares method is used to construct the first regression model to obtain the roasted porosity prediction model of single iron ore powder.

[0012] S4. Obtain the blending ratio of the mixed ore in the actual sintering production, calculate the predicted roasting porosity of each single type of iron ore powder according to the prediction model described in step S3, and calculate the theoretical porosity of the blended ore after roasting by weighting according to the blending ratio.

[0013] S5. Obtain multiple sets of actual production data, each set of data including: actual porosity of sinter, theoretical porosity of the blended ore after roasting corresponding to the sinter, and sintering process parameters corresponding to the production of the sinter.

[0014] S6. Using the theoretical porosity of the roasted mixed ore and the sintering process parameters as independent variables and the actual porosity of the sintered ore as the dependent variable, a second regression model is constructed using the partial least squares method to obtain the final prediction model of the porosity of the sintered ore in actual production.

[0015] S7. Input the target ore blending scheme and target sintering process parameters into the final prediction model to obtain the predicted value of sinter porosity, and use this to guide the adjustment of ore blending and / or sintering process parameters.

[0016] Preferably, the prediction model for the roasting porosity of single-variety iron ore powder described in step S3 is as follows:

[0017]

[0018] Among them, K d The predicted porosity of single-variety iron ore powder after roasting is given. A, B, C, D, E, F, G, H, I, and J represent the mass fractions of hematite, quartz, TFe content, and limonite, respectively, after standardization. Content, magnetite mass fraction, MgO content, CaO content, Content, FeO content.

[0019] Preferably, the theoretical porosity K of the roasted mixed ore described in step S4 is... hl Calculated using the following formula:

[0020]

[0021] Among them, K d For the predicted roasting porosity of each single type of iron ore powder, P i This refers to the proportion of a single type of iron ore powder in a blended ore.

[0022] Preferably, for step S2, the fixed binary basicity R is 2.0; the conditions for laboratory roasting are: roasting temperature 1280℃, constant temperature time 4min.

[0023] Preferably, for step S5, the sintering process parameters include at least: sinter R value, solid fuel ratio, limestone ratio, quicklime ratio, and dolomite ratio.

[0024] Preferably, the final prediction model described in step S6 is:

[0025]

[0026] Among them, K s K represents the predicted actual porosity of sinter. hl Theoretical porosity after roasting of the homogenized ore is given by: R, S, L, Q, D. Theoretical porosity of the sintered ore is given by: S, L, Q, D. Theoretical porosity of the solid fuel is given by: L, Q, D. Theoretical porosity of the limestone is given by: Q, D. Theoretical porosity of the dolomite is given by:

[0027] Preferably, step S7 further includes:

[0028] Set the control threshold range for the porosity of sintered ore;

[0029] When the predicted value exceeds the threshold range, the predicted value is brought within the threshold range by adjusting the ore blending structure to change the theoretical porosity of the blended ore and / or adjusting the sintering process parameters.

[0030] A porosity prediction system based on iron ore powder and mineral composition information is used to implement a porosity prediction method based on iron ore powder and mineral composition information, including:

[0031] The data acquisition module is configured to acquire and store the chemical composition and mineral composition data of a single type of iron ore powder, the blending ratio data of the mixed ore, the sintering process parameter data, and the measured porosity data of the sintered ore.

[0032] The first prediction model construction module is configured as follows: using the main chemical component content and main mineral composition mass fraction of each single iron ore powder as independent variables, and the corresponding roasted porosity as dependent variable, the partial least squares method is used to construct the first regression model and construct the single iron ore powder roasted porosity prediction model.

[0033] The porosity calculation module is configured to: obtain the blending ratio of the mixed ore in the actual sintering production, and calculate the theoretical porosity of the roasted mixed ore based on the first prediction model and the blending ratio.

[0034] The second prediction model construction module is configured to: acquire multiple sets of actual production data, use partial least squares method to construct a second regression model, and obtain the final prediction model of sinter porosity during actual production.

[0035] The porosity prediction and optimization module is configured to: output the predicted porosity value using the final prediction model, and provide suggestions on the direction of adjustment of ore blending and / or process parameters based on the model parameters when the predicted value does not meet the set threshold.

[0036] Preferably, the system further includes a model iteration and update module, configured to: add newly added single-variety iron ore powder data and production data to the training set, and retrain and update the first prediction model and / or the second prediction model.

[0037] Compared with the prior art, the beneficial effects of the present invention are:

[0038] 1. This invention constructs a sintering behavior prediction model based on the chemical composition and mineral structure of a single iron ore powder. It correlates the compositional information of each iron ore powder with its roasting porosity under standard conditions and establishes an accurate mathematical regression model using partial least squares (PLS) to achieve a quantitative assessment of the sintering porosity tendency of a single raw material. This allows for advance understanding of the contribution direction and degree of different iron ore powders to porosity, enabling raw material screening and optimization during the ore blending stage. In actual production, the ore blending structure and some sintering process parameters can be optimized and adjusted in a timely manner based on the prediction results, ensuring that the porosity of the sintered ore remains within a reasonable range during actual production. This provides significant guidance for sintering quality improvement and effectively supports blast furnace production.

[0039] 2. This invention proposes the variable of theoretical porosity of blended ore and establishes a final porosity prediction model that integrates raw material characteristics and real-time process. Through single-ore model prediction and weighted calculation, the complex ore blending scheme is organized into a theoretical porosity of blended ore that characterizes the comprehensive characteristics of raw materials. This theoretical value, together with the key parameters of the actual sintering process, is used as input and modeled again with the measured final porosity of sintered ore. This enables accurate prediction of the final porosity of sintered ore under given ore blending and process parameters. Attached Figure Description

[0040] Figure 1 This is a schematic diagram of the prediction method of the present invention. Detailed Implementation

[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0042] To address the issue that traditional sintering methods cannot accurately and quantitatively predict the porosity of sintered ore before production, please refer to [link / reference needed]. Figure 1 This embodiment provides the following technical solution:

[0043] A porosity prediction method based on iron ore powder and mineral composition information, the main prediction process includes:

[0044] The main chemical composition and mineral composition of a number of single-variety iron ore powders were obtained through experiments. At the same time, CaO reagent was added to the above-mentioned single-variety iron ore powders at a fixed R value, and the powders were pressed into sample cakes and roasted. The porosity of the roasted samples was measured. The porosity of the roasted single-variety iron ore powders obtained by the test was analyzed with the corresponding chemical composition and main mineral composition of the iron ore powders using partial least squares data analysis to determine the mathematical model for predicting the porosity of sintered single-variety iron ore powders.

[0045] The ore blending ratio of the sintered homogenized ore during actual sintering production, as well as sintering process parameters (including but not limited to the R value of sintered ore, the ratio of sintered solid fuel, the ratio of quicklime, the ratio of limestone, and the ratio of dolomite), are obtained. The porosity of the sintered homogenized ore after roasting is calculated by weighting the predicted value of the roasted porosity of the single-variety iron ore powder with the corresponding blending ratio of homogenized ore. At the same time, the actual porosity of the sintered ore during actual sintering production is obtained through experiments.

[0046] By performing partial least squares data analysis on the actual porosity of sinter during the above-mentioned sintering production, the porosity calculated after roasting the corresponding blended ore, and the corresponding sintering process parameters (including but not limited to the R value of sinter, the ratio of sintering solid fuel, the ratio of quicklime, the ratio of limestone, and the ratio of dolomite), a predictive mathematical model for the porosity of sinter during actual production is finally obtained, thereby guiding the optimization of ore blending and the reasonable adjustment of sintering process parameters.

[0047] The detailed implementation steps are illustrated in the following example:

[0048] S1. The main chemical components of various single iron ore powders to be sintered are obtained by laboratory analysis, including but not limited to TFe, CaO , MgO, FeO, as shown in Table 1.

[0049] Table 1. Main chemical components of various iron ore powders

[0050]

[0051] S2. Prepare thin sections of each of the above-mentioned single iron ore powders, and quantitatively analyze the mineral composition of the iron ore powders using the through-scale method under a dual-purpose polarized light microscope (transmitting / reflecting light) to obtain the mass fractions of magnetite, hematite, limonite, and quartz in the mineral composition of each iron ore powder. The results are shown in Table 2.

[0052] Table 2 Main mineral composition of each type of iron ore powder

[0053]

[0054] S3. Take samples of each type of iron ore powder separately, dry them, and grind them using a grinder until the particle size is less than 0.074 mm. Add CaO analytical grade reagent to adjust R2 (binary basicity) to 2.0 and mix thoroughly. Take 0.8 g of the mixture and put it into a manual sample press. Press the mixture into small cakes of 8 mm × 5 mm at 15 MPa. Place the small cakes into a horizontal tube furnace for roasting. Set the roasting temperature to 1280 ℃ and the constant temperature time to 4 min.

[0055] S4. Prepare thin sections from the roasted samples described above, and determine the porosity of the roasted iron ore powder using a polarizing microscope with both transmission and reflection capabilities and the method of measuring through-scale. Specific data are shown in Table 3.

[0056] Table 3 Porosity of individual iron ore powders after roasting

[0057]

[0058] S5. Using the porosity of each individual iron ore powder after roasting, the corresponding main chemical composition of the iron ore powder, and the mass fraction of the main mineral composition of each individual iron ore powder as independent variables, and the porosity of the corresponding iron ore powder after roasting as the dependent variable, a regression mathematical model is constructed using partial least squares method. The regression model coefficients and R0 are... 2 As shown in Table 4.

[0059] Table 4 Regression model coefficients and R 2

[0060]

[0061] The specific regression model is shown in Equation 1:

[0062] Formula 1: Porosity K of a single type of iron ore powder after roasting d = 27.883 + 1.96 * Hematite (standardized) + 0.131 * Quartz (standardized) - 0.653 * TFe (standardized) + 1.484 * Limonite (standardized) + 1.404 * (Standardized) -3.17 * Magnetite (Standardized) +0.544 * MgO (Standardized) +0.388 * CaO (Standardized) -0.946 * (Standardized) +2.504* FeO (Standardized)

[0063] S6. During actual sintering production, the proportion of each type of iron ore powder participating in sintering is obtained, and the theoretical porosity of the mixed ore is calculated. Specifically, it is obtained by weighting the proportion of each type of ore and the porosity of the corresponding iron ore powder after roasting. The binder strength of the individual iron ore powder is calculated according to Formula 1, as shown in Formula 2 below:

[0064]

[0065] Where: K hl The theoretical porosity of the blended ore after roasting is expressed in %; K d Porosity of each single type of iron ore powder after roasting, in %; P i The proportion of a single type of iron ore powder is given, in units of %.

[0066] S7. Obtain several sets of sintered ore samples from actual production, prepare thin sections, and determine the porosity of the roasted iron ore powder using a polarizing microscope with both transmission and reflection capabilities and the over-scale measurement method. Obtain the main chemical composition, main mineral composition, and sintering ratio of each individual iron ore powder. Calculate the porosity of each individual iron ore powder after roasting according to Formula 1, and calculate the theoretical porosity of the corresponding blended ore after roasting according to Formula 2. Simultaneously, obtain the sintering process parameters corresponding to the time sequence of the sintered ore samples, such as the R value of the sintered ore, the ratio of solid fuel for sintering, the ratio of quicklime, the ratio of limestone, and the ratio of dolomite. Detailed data are shown in Table 5.

[0067] Table 5 Actual porosity of sintered ore, theoretical porosity of blended ore, and process parameters

[0068]

[0069] S8. Using the theoretical porosity of the roasted mixed ore, the sinter R, the solid fuel ratio, the limestone ratio, the quicklime ratio, and the dolomite ratio as independent variables, and the corresponding actual porosity Ks of the sinter as the dependent variable, a regression mathematical model was constructed using partial least squares method. The regression model coefficients and R² are shown in Table 6.

[0070] Table 6 Regression Model Coefficients and R 2

[0071]

[0072] The final prediction model is shown in Formula 3 below:

[0073] Formula 3: Actual porosity of sintered ore Ks = 1.02 * Theoretical porosity after roasting of blended ore - 0.15 * Sintered ore R - 0.11 * Solid fuel ratio + 0.27 * Limestone ratio - 0.08 * Quicklime ratio + 0.07 * Dolomite ratio - 0.31

[0074] S9. If new data on the main chemical composition and mineral composition of single-variety iron ore powder, as well as new data on the actual porosity of sintered ore, are generated, these data will be added to the basic database as new data sources. Partial least squares iterative calculations will then be performed to obtain new predictive regression models. Furthermore, as the amount of data increases, the predictive accuracy improves, thus providing significant guidance for actual sintering blending and production.

[0075] In actual sintering production, when the sinter quality (R and FeO) is relatively stable within a controlled range, a control threshold for sinter porosity is set. This is primarily achieved through ore blending optimization, supplemented by adjustments to the limestone-quicklime ratio, to ensure the sinter porosity remains within a reasonable threshold range. When the sinter R is controlled within the range of 1.80-2.10, FeO within the range of 8.0-9.0, and MgO within the range of 1.85-2.05, and the actual porosity control threshold is between 26% and 32%, the sinter exhibits superior strength and low-temperature reduction pulverization properties.

[0076] Based on the porosity prediction method described above, two specific prediction examples are provided below:

[0077] Case 1: The limestone ratio is 2.5%, quicklime ratio is 3.2%, fuel ratio is 4%, dolomite ratio is 5%, and sinter R is 2.0. Under a certain ore blending structure, the porosity of the roasted mixed ore is 25%. According to the above prediction model, the final porosity of the sinter is predicted to be 25.22%, which is lower than the lower limit of the sinter porosity control threshold. At this time, the ore blending is optimized and adjusted so that the porosity of the roasted mixed ore reaches 25.80%. The actual porosity of the corresponding sinter reaches 26.04%, which meets the control threshold range requirement.

[0078] Case 2: When the limestone ratio is 2.5%, the quicklime ratio is 3.2%, the fuel ratio is 4%, the dolomite ratio is 5%, and the sinter R is 2.0, the porosity of the blended ore after roasting under a certain ore blending structure is 32%. According to the above prediction model, the final sinter porosity is predicted to be 32.36%, which is higher than the upper limit of the sinter porosity control threshold. At this time, the original ore blending structure can be maintained. While keeping the sinter R consistent, production can be carried out with a full quicklime ratio. Then the limestone ratio is 0, and the quicklime ratio is adjusted to 4.8%. According to the above prediction model, the actual sinter porosity is 31.56%, which meets the control threshold range requirement.

[0079] Working principle: First, detailed information on various iron ore powders, including key chemical components and mineral composition, is obtained. Micro-sintering experiments are then conducted on these single ore powders under strictly controlled alkalinity, temperature, and time conditions to measure the standard roasted porosity. Subsequently, partial least squares (PLS) multivariate statistical analysis is used to establish a regression equation between the ore powder information and the roasted porosity. The coefficients of this equation quantitatively reveal the direction and degree of contribution of each component and mineral to porosity formation; for example, hematite and limonite content are positively correlated with porosity, while magnetite content is negatively correlated.

[0080] In actual sintering, a blend of various iron ore powders in specific proportions is used, along with fuel, flux, and other additives, and the process parameters are constantly changing. Therefore, a first-level model is first used to predict the standard porosity of each individual ore powder based on its characteristics. Then, a weighted average is calculated according to the actual blending ratio to determine the theoretical porosity of the blend under standard conditions. This value represents the porosity tendency of the current blend structure itself. Next, a large number of actual production data sets are collected. Each set includes: the theoretical porosity of the blend calculated from the blending, the key process parameters during the production of that batch, and the actual porosity of the sinter measured for that batch. Partial least squares is then applied again to establish a final prediction model with the theoretical porosity of the blend and various process parameters as independent variables, and the actual porosity of the sinter as the dependent variable.

[0081] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0082] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A porosity prediction method based on iron ore powder and mineral composition information, characterized in that, Includes the following steps: S1. Obtain the content of major chemical components and the mass fraction of major mineral composition of various single-variety iron ore powders. The major chemical components include at least TFe. CaO The main mineral composition includes at least magnetite, hematite, limonite, and quartz, and consists of MgO and FeO. S2. Each type of iron ore powder was roasted in the laboratory under a fixed binary basicity R, and its porosity after roasting was measured. S3. Using the main chemical components and the mass fraction of the main mineral composition of each single iron ore powder as independent variables and the corresponding roasted porosity as dependent variable, the partial least squares method is used to construct the first regression model to obtain the roasted porosity prediction model of single iron ore powder. S4. Obtain the blending ratio of the mixed ore in the actual sintering production, calculate the predicted roasting porosity of each single type of iron ore powder according to the prediction model described in step S3, and calculate the theoretical porosity of the blended ore after roasting by weighting according to the blending ratio. S5. Obtain multiple sets of actual production data, each set of data including: actual porosity of sinter, theoretical porosity of the blended ore after roasting corresponding to the sinter, and sintering process parameters corresponding to the production of the sinter. S6. Using the theoretical porosity of the roasted mixed ore and the sintering process parameters as independent variables and the actual porosity of the sintered ore as the dependent variable, a second regression model is constructed using the partial least squares method to obtain the final prediction model of the porosity of the sintered ore in actual production. S7. Input the target ore blending scheme and target sintering process parameters into the final prediction model to obtain the predicted value of sinter porosity, and use this to guide the adjustment of ore blending and / or sintering process parameters.

2. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, The prediction model for the porosity of roasted single-variety iron ore powder mentioned in step S3 is as follows: Among them, K d The predicted porosity of single-variety iron ore powder after roasting is given. A, B, C, D, E, F, G, H, I, and J represent the mass fractions of hematite, quartz, TFe content, and limonite, respectively, after standardization. Content, magnetite mass fraction, MgO content, CaO content, Content, FeO content.

3. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, Regarding the theoretical porosity K of the roasted mixed ore described in step S4 hl Calculated using the following formula: Among them, K d For the predicted roasting porosity of each single type of iron ore powder, P i This refers to the proportion of a single type of iron ore powder in a blended ore.

4. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, For step S2, the fixed binary basicity R is 2.0; the conditions for laboratory calcination are: calcination temperature 1280℃, constant temperature time 4min.

5. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, For step S5, the sintering process parameters include at least: sinter R value, solid fuel ratio, limestone ratio, quicklime ratio, and dolomite ratio.

6. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, The final prediction model described in step S6 is as follows: Among them, K s K represents the predicted actual porosity of sinter. hl Theoretical porosity after roasting of the homogenized ore is given by: R, S, L, Q, D. Theoretical porosity of the sintered ore is given by: S, L, Q, D. Theoretical porosity of the solid fuel is given by: L, Q, D. Theoretical porosity of the limestone is given by: Q, D. Theoretical porosity of the dolomite is given by:

7. The porosity prediction method based on iron ore powder and mineral composition information according to claim 1, characterized in that, Step S7 further includes: Set the control threshold range for the porosity of sintered ore; When the predicted value exceeds the threshold range, the predicted value is brought within the threshold range by adjusting the ore blending structure to change the theoretical porosity of the blended ore and / or adjusting the sintering process parameters.

8. A porosity prediction system based on iron ore powder and mineral composition information, used to implement the porosity prediction method based on iron ore powder and mineral composition information as described in any one of claims 1-7, characterized in that, include: The data acquisition module is configured to acquire and store the chemical composition and mineral composition data of a single type of iron ore powder, the blending ratio data of the mixed ore, the sintering process parameter data, and the measured porosity data of the sintered ore. The first prediction model construction module is configured as follows: using the main chemical component content and main mineral composition mass fraction of each single iron ore powder as independent variables, and the corresponding roasted porosity as dependent variable, the partial least squares method is used to construct the first regression model and construct the single iron ore powder roasted porosity prediction model. The porosity calculation module is configured to: obtain the blending ratio of the mixed ore in the actual sintering production, and calculate the theoretical porosity of the roasted mixed ore based on the first prediction model and the blending ratio. The second prediction model construction module is configured to: acquire multiple sets of actual production data, use partial least squares method to construct a second regression model, and obtain the final prediction model of sinter porosity during actual production. The porosity prediction and optimization module is configured to: output the predicted porosity value using the final prediction model, and provide suggestions on the direction of adjustment of ore blending and / or process parameters based on the model parameters when the predicted value does not meet the set threshold.

9. A porosity prediction system based on iron ore powder and mineral composition information according to claim 8, characterized in that, The system also includes a model iteration and update module, configured to add newly added single-variety iron ore powder data and production data to the training set, and retrain and update the first prediction model and / or the second prediction model.

Images