Method and system for determining clean coal yield

Abstract

The invention discloses a clean coal yield determination method and system, and relates to the technical field of coal reselection and flotation, and the clean coal yield determination method comprises the following steps: collecting historical floating and sinking data of reselection feeding materials and distribution release data of flotation feeding materials; predicting the ash content of the reselection feeding material and the ash content of the flotation feeding material according to the ash content of the raw coal; correcting historical floating and sinking data to generate a new floating and sinking table; fitting and drawing a gravity separation actual selectability curve; correcting the yield of the historical step-by-step release data table, fitting and drawing an actual floatability curve; reselection process parameters are determined, and flotation process parameters are determined; and determining a product structure, and setting corresponding production parameters. The method can improve the clean coal yield.

Description

Technical Field

[0001] The invention relates to the technical field of coal re-selection and flotation, and in particular to a method and system for determining clean coal yield. Background Art

[0002] In the process of coal washing, gravity separation refers to a method of separation of mineral particles by using the differences in relative density, particle size, shape and movement speed and direction in the medium (water, air or other liquids with high relative density). Flotation is a method of separation of useful minerals from useless minerals based on the differences in the surface properties of minerals under the action of reagents. The composition of comprehensive clean coal in a coal preparation plant usually includes gravity clean coal, coarse clean coal slime and flotation clean coal. The combination of clean coal ash in each link is different, and comprehensive clean coal with ash content that meets user requirements can still be obtained. However, different combinations of clean coal ash will result in different comprehensive clean coal yields.

[0003] In the existing technology, most coal preparation plants use the following washing process: heavy medium for lump and fine coal - separate recovery of coarse coal slime - coal slime flotation process. In the actual production process, in order to ensure that the ash content of the comprehensive clean coal meets the standard, each workshop formulates the clean coal ash content index of each link based on manual experience, thereby determining the separation parameters of each link. The formulated clean coal ash content index of each link is usually far lower than the sales standard, thus sacrificing the clean coal yield to ensure that the clean coal ash content does not exceed the standard.

[0004] At present, the existing technology cannot accurately obtain the clean coal ash content that does not exceed the standard based on manual experience, so it is impossible to reasonably set the production parameters, and unreasonable production parameters will affect the clean coal yield. Summary of the invention

[0005] The embodiments of the present invention provide a method and system for determining the clean coal yield, which can solve the problem in the prior art that the clean coal yield cannot be improved by reasonably setting production parameters through accurate clean coal ash content.

[0006] The embodiment of the present invention provides a method for determining the clean coal yield, comprising the following steps: obtaining a float-sink table containing a float-sink table yield and a historical gravity separation feed ash content under gravity separation, and a step-by-step release data table containing a step-by-step release data table yield and a historical flotation feed ash content under flotation separation; the gravity separation feed ash content represents the content of non-combustible components obtained after mineral particles in the raw coal are separated by gravity separation, and the flotation feed ash content represents the content of non-combustible components obtained after mineral particles in the raw coal are separated by flotation; using the raw coal ash content to train a neural network model to predict the gravity separation feed ash content and the flotation feed ash content, and obtaining an ash prediction model; inputting the raw coal ash to be detected into the ash prediction model, and obtaining the gravity separation feed ash content and the flotation feed ash content. ash content of the material; according to the difference between the ash content of the gravity separation material and the historical gravity separation material ash content, the float-sink table yield is corrected and the float-sink table is updated; according to the difference between the ash content of the flotation material and the historical flotation material ash content, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated float-sink table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted; according to the boundary ash content used to distinguish the clean coal, the gravity separation clean coal ash content and the flotation clean coal ash content are obtained in the gravity separation actual selectivity curve and the actual floatability curve respectively; according to the gravity separation clean coal ash content and the flotation clean coal ash content, the equal boundary ash content principle is used to obtain the total clean coal ash content, and the corresponding production parameters are set according to the total clean coal ash content to determine the clean coal yield.

[0007] Furthermore, the specific steps of correcting the yield of the float-sink table include:

[0008] The difference Δweight between the ash content of the gravity separation feed and the ash content of the historical gravity separation feed is obtained by the formula:

[0009] Δ reentry = λ reentry - λ reentry

[0010] Wherein: λre-entry is the ash content of the gravity feed; λre-original is the ash content of the historical gravity feed, and the historical gravity feed ash content is specifically: the subtotal ash in the historical gravity feed float and sink table;

[0011] Get density greater than +1.8g / cm 3 The density level of ash Ad +1.8 , the formula is;

[0012]

[0013] Where Adn means the nth one is greater than +1.8g / cm 3 The density level of ash; γn means the nth density level greater than +1.8g / cm 3 The yield of density grade; n means greater than +1.8g / cm 3 The number of density levels;

[0014] Get density less than -1.8g / cm 3 The density level of ash Ad -1.8 , the formula is;

[0015]

[0016] Among them, Adm means the mth one is less than -1.8g / cm 3 The density level of ash; γm means the mth one is less than -1.8g / cm 3 The yield of density grade; m means less than -1.8g / cm 3 The number of density levels;

[0017] Get the calibration value X, the formula is:

[0018] X = (100 * Δ weight) / (Ad +1.8 -Ad -1.8 )

[0019] Correction of the yield of the float-sink table;

[0020] If the density level is +1.8g / cm 3 :

[0021] γ correction = γ + X

[0022] If the density level is -1.8g / cm 3 :

[0023]

[0024] Here, γ correction represents the yield of the corresponding density level after correction, and γ represents the yield of the corresponding density level before correction.

[0025] Furthermore, the correction of the yield of the distribution release data table comprises the following specific steps:

[0026] The difference Δfloat between the ash content of the flotation feed and the ash content of the historical flotation feed is obtained. The formula is:

[0027] Δ float = λ float in - λ float out

[0028] Among them, λ float in is the flotation feed ash; λ float original is the historical flotation feed ash, and the historical flotation feed ash is specifically: the weighted ash of each clean tailing coal in the historical distribution release data table;

[0029] Get distribution release of the first tail coal ash Ad 轻 , the formula is;

[0030]

[0031] Among them, Adtailm represents the ash content of the mth tail coal released by distribution; γtailm represents the yield of the mth tail coal released by distribution; m represents the number of flotation in the distribution release test; Adfine represents the ash content of the clean coal released by distribution; γfine represents the yield of the clean coal released by distribution; γtail1 represents the yield of the first tail coal released by distribution;

[0032] Get the calibration value X, the formula is:

[0033] X = (100 * Δ float) / (Ad 尾1 -Ad 轻 )

[0034] Among them, Ad 尾1 Indicates the distribution and release of the ash content of the first tail coal;

[0035] The distribution release table yield was corrected;

[0036] For distribution and release of first flotation tailings:

[0037] γ correction = γ tail 1 + X

[0038] For the distribution release of the mth flotation tailings (m>=2):

[0039] γ correction = γ tail m-γ tail m*X / (100-γ tail 1)

[0040] For distribution release of clean coal:

[0041] γ correction = γ essence - γ essence * X / (100-γ tail 1)

[0042] Among them, γ correction represents the yield of clean (tail) coal at the corresponding flotation times after correction.

[0043] Furthermore, the ash content of the gravity-separated clean coal is obtained in the actual selectivity curve of gravity separation, and the specific steps include: setting a boundary ash content for distinguishing the clean coal according to the target ash content required as the ash content of the clean coal; on the ash characteristic curve within the actual selectivity curve of gravity separation, the boundary ash content is used as the horizontal coordinate value, and the corresponding vertical coordinate value is obtained as the gravity-separated clean coal yield; the actual selectivity curve of gravity separation includes: an ash characteristic curve λ, a floating matter accumulation curve β, and a density curve δ; on the floating matter accumulation curve within the actual selectivity curve of gravity separation, the gravity-separated clean coal yield is used as the vertical coordinate value, and the corresponding horizontal coordinate value is obtained as the gravity-separated clean coal ash content.

[0044] Further, the steps for obtaining the ash content of the flotation clean coal in the actual floatability curve are as follows: on the ash content characteristic curve within the actual floatability curve, using the boundary ash content as the abscissa value, obtain the corresponding ordinate value as the yield of the flotation clean coal; the actual floatability curve includes: the ash content characteristic curve λ, the cumulative float curve β, the cumulative float yield - distribution release test times curve v, and the cumulative sink curve n; on the cumulative float curve within the actual floatability curve, using the yield of the flotation clean coal as the ordinate value, obtain the corresponding abscissa value as the ash content of the flotation clean coal.

[0045] Further, the steps for obtaining the total clean coal ash content using the equal boundary ash content principle are as follows:

[0046] Obtain the total clean coal ash content A_calculated according to the weighted average ash content of the heavy - selected clean coal ash content and the flotation clean coal ash content;

[0047] According to the equal boundary ash content principle, when the difference between the total clean coal ash content and the target clean coal ash content is less than the set error standard value, the clean coal yield obtained is higher than that when the difference between the total clean coal ash content and the target clean coal ash content is greater than the set error standard value;

[0048] Set the error standard value ε, the step size α, and the target clean coal ash A_target;

[0049] If |A_calculated - A_target| > ε:

[0050] When A_calculated > A_target, A_calculated = A_calculated - α; when A_calculated < A_target, A_calculated = A_calculated + α;

[0051] If |A_calculated - A_target| < ε, the total clean coal ash content A_calculated meets the requirements.

[0052] Further, the steps for setting corresponding production parameters based on the total clean coal ash content are as follows:

[0053] On the density curve within the actual可选性曲线 of gravity separation, using the total clean coal ash content as the ordinate value, obtain the corresponding abscissa value as the separation density of the gravity separation section; on the cumulative float yield - distribution release test times curve within the actual floatability curve, using the total clean coal ash content as the ordinate value, obtain the corresponding abscissa value as the optimal flotation stage number of the flotation section.

[0054] An embodiment of the present invention provides a determination system for the clean coal yield, including:

[0055] An ash content prediction module, configured to obtain a float - sink table including the float - sink table yield and the historical heavy - selected feed ash content in the case of gravity separation, and a step - by - step release data table including the step - by - step release data table yield and the historical flotation feed ash content in the case of flotation;

[0056] The yield correction module is used to use the raw coal ash content to train the neural network model to predict the gravity separation feed ash content and the flotation feed ash content to obtain the ash content prediction model; input the raw coal ash content to be detected into the ash content prediction model to obtain the gravity separation feed ash content and the flotation feed ash content; according to the difference between the gravity separation feed ash content and the historical gravity separation feed ash content, the float-sink table yield is corrected and the float-sink table is updated; according to the difference between the flotation feed ash content and the historical flotation feed ash content, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated float-sink table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted;

[0057] The yield improvement module is used to obtain the gravity separation clean coal ash content and the flotation clean coal ash content in the actual selectivity curve and the actual floatability curve of gravity separation according to the boundary ash content used to distinguish the clean coal; according to the gravity separation clean coal ash content and the flotation clean coal ash content, the total clean coal ash content is obtained using the equal boundary ash principle, and the corresponding production parameters are set according to the total clean coal ash content to determine the clean coal yield.

[0058] The embodiment of the present invention provides a method and system for determining the clean coal yield, which has the following beneficial effects compared with the prior art:

[0059] The ash content of gravity separation feed and flotation feed are obtained through the trained neural network model - ash prediction model. According to the difference between the gravity separation feed ash and the historical gravity separation feed ash, the floating and sinking table yield is corrected and the floating and sinking table is updated; according to the difference between the flotation feed ash and the historical flotation feed ash, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated floating and sinking table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted; the gravity separation clean coal ash and flotation clean coal ash are obtained through the corresponding values ​​of the curves, and the total clean coal ash is obtained through the equal boundary ash principle, and finally the corresponding production parameters are determined according to the total clean coal ash. While avoiding the error caused by obtaining the total clean coal ash by artificial experience, the production parameters can be accurately set by the total clean coal ash, so as to achieve the effect of improving the clean coal yield according to the accurately set production parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of a process flow provided by an embodiment of the present invention;

[0061] Figure 2 A corrected floating and sinking data table provided in an embodiment of the present invention;

[0062] Figure 3 The actual selectivity curve diagram of gravity separation drawn after fitting provided by the embodiment of the present invention, wherein the λ curve represents the ash characteristic curve; the β curve represents the floating matter accumulation curve; the δ curve represents the density curve;

[0063] Figure 4 A corrected step-by-step release data table provided in an embodiment of the present invention;

[0064] Figure 5 The actual floatability curve diagram drawn after fitting provided by the embodiment of the present invention, wherein the λ curve represents the ash characteristic curve; the β curve represents the floating matter accumulation curve; the v curve represents the floating matter accumulation yield-distribution release test number curve; and the n curve represents the sinking matter accumulation curve. DETAILED DESCRIPTION

[0065] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without violating the connotation of the present invention, so the present invention is not limited by the specific embodiments disclosed below.

[0066] See also Figure 1 The embodiment of the present invention provides a method for determining the clean coal yield, comprising the following steps:

[0067] Step 1: Obtain a float-sink table containing the float-sink table yield and the historical gravity separation feed ash content in the case of gravity separation, and a step-by-step release data table containing the step-by-step release data table yield and the historical flotation feed ash content in the case of flotation.

[0068] Step 2: Use the raw coal ash content and raw coal screening data to train the neural network model to predict the gravity separation feed ash content and the flotation feed ash content to obtain the ash content prediction model; input the raw coal ash content to be tested into the ash content prediction model to obtain the gravity separation feed ash content and the flotation feed ash content; according to the difference between the gravity separation feed ash content and the historical gravity separation feed ash content, correct the floatation and sinking table yield and update the floatation and sinking table; according to the difference between the flotation feed ash content and the historical flotation feed ash content, correct the distribution release data table yield and update the distribution release data table; according to the updated floatation and sinking table, fit the actual selectivity curve of gravity separation; according to the updated distribution release data table, fit the actual floatability curve.

[0069] Step three: According to the boundary ash content used to distinguish clean coal, obtain the gravity separation clean coal ash content and flotation clean coal ash content in the actual selectivity curve and actual floatability curve of gravity separation respectively; according to the gravity separation clean coal ash content and the flotation clean coal ash content, use the equal boundary ash content principle to obtain the total clean coal ash content, and set the corresponding production parameters according to the total clean coal ash content to determine the clean coal yield.

[0070] The detailed explanation of the above steps is as follows:

[0071] 1. Based on the raw coal ash content and raw coal screening results, the BP neural network model is trained to determine the density composition of the coal, thereby predicting the ash content of the gravity separation feed and the flotation feed.

[0072] 2. The historical floating and sinking data of the gravity separation feed and the floatability data (distribution release data) of the flotation feed are collected through the coal preparation plant centralized control system.

[0073] 3. Correction of historical floating and sinking data: Input the new ash content of the feed material, and calculate the difference between the feed ash content and the original table ash content Δweight. Use Δweight to correct the yield of the original floating and sinking table, and generate a new floating and sinking table based on the corrected yield and ash content, such as Figure 2 shown.

[0074] The specific steps include:

[0075] 3.1 Calculate Δ weight:

[0076] Δ reentry = λ reentry - λ reentry

[0077] Among them: λ re-entry is the ash content of the gravity-selected material; λ gravity original is the ash content of the historical gravity-selected material. The specific historical gravity-selected material ash content is: the subtotal ash content in the historical gravity-selected material floating and sinking table.

[0078] 3.2 Calculation of Ad +1.8 :

[0079]

[0080] Where Adn means the nth one is greater than +1.8g / cm 3 The density level of ash; γn means the nth density level greater than +1.8g / cm 3 The yield of density grade; n means greater than +1.8g / cm 3 The number of density levels.

[0081] 3.3 Calculation of Ad -1.8 :

[0082]

[0083] Where: Adm means the mth one is less than -1.8g / cm 3 The density level of ash; γm means the mth one is less than -1.8g / cm 3 The yield of density grade; m means less than -1.8g / cm 3 The number of density levels.

[0084] 3.4 Calculate the calibration value X:

[0085] X = (100 * Δ weight) / (Ad +1.8 -Ad -1.8 )

[0086] 3.5 Correction of the yield of the floating and sinking table:

[0087] If the density level is +1.8g / cm 3 :

[0088] γ correction = γ + X

[0089] If the density level is -1.8g / cm 3 :

[0090]

[0091] Here, γ correction represents the yield of the corresponding density level after correction, and γ represents the yield of the corresponding density level before correction.

[0092] 4. Plotting of actual selectivity curve: The percentage of a certain density level in the raw material that enters the light product is called the distribution rate of the density level in the light product, which is represented by the symbol e. Calculate the corresponding distribution rates of the five values ​​of sorting density of 1.3, 1.4, 1.5, 1.6 and 1.7: find the value of the middle value t, calculate the value of e according to the high-level language python, calculate the data table of the actual selectivity curve of gravity sorting, fit and plot the actual selectivity curve of gravity sorting, such as Figure 3 shown.

[0093] The empirical formula for finding the intermediate parameter t is:

[0094] t=0.675 / E*(δ-δp)

[0095] Among them, δ represents the average density of each density level, and δp represents the actual sorting density.

[0096] The Python high-level language for finding the value of e is:

[0097]

[0098] The fitting formula of the actual floatability curve is:

[0099] y=100*(b-np.arctan(c*(xd)))*(ba)

[0100] Among them, a, b, c, d, and x are parameters to be determined.

[0101] 5. Correction of historical step-by-step release data: Input the new flotation feed ash content, and calculate the difference between the feed ash content and the total ash content in the original step-by-step release data table, Δfloat. Using Δfloat, correct the yield of the original step-by-step release data table, and generate a new distribution release data table based on the corrected yield and ash content, such as Figure 4 As shown, the actual floatability curve is fitted and plotted, as shown in Figure 5shown.

[0102] The difference Δfloat between the ash content of the flotation feed and the ash content of the historical flotation feed is obtained. The formula is:

[0103] Δ float = λ float in - λ float out

[0104] Among them, λ float in is the flotation feed ash; λ float original is the historical flotation feed ash, and the historical flotation feed ash is specifically: the weighted ash of each clean tailing coal in the historical distribution release data table.

[0105] Get distribution release of the first tail coal ash Ad 轻 , the formula is;

[0106]

[0107] Among them, Ad tail m represents the ash content of the mth tail coal released by distribution; γ tail m represents the yield of the mth tail coal released by distribution; m represents the number of flotation times in the distribution release test; Ad fine represents the ash content of the distributed released clean coal; γ fine represents the yield of the distributed released clean coal; γ tail 1 represents the yield of the first tail coal released by distribution.

[0108] Get the calibration value X, the formula is:

[0109] X = (100 * Δ float) / (Ad 尾1 -Ad 轻 )

[0110] Among them, Ad 尾1 Indicates the distribution and release of the ash of the first tail coal.

[0111] Yields were corrected for distribution release.

[0112] For distribution and release of first flotation tailings:

[0113] γ correction = γ tail 1 + X

[0114] For the distribution release of the mth flotation tailings (m>=2):

[0115] γ correction = γ tail m-γ tail m*X / (100-γ tail 1)

[0116] For distribution release of clean coal:

[0117] γ correction = γ essence - γ essence * X / (100-γ tail 1)

[0118] Among them, γ correction represents the yield of clean (tail) coal of the corresponding flotation times after correction, and γ* represents the yield of clean (tail) coal of the corresponding flotation times before correction.

[0119] 6. Determine the re - washing process parameters: Set the boundary ash content λ according to the target ash content \(A_{target}\). Check the ash content characteristic curve (λ curve) of the actual washability curve of re - washing, so as to obtain the corresponding ordinate \(a_2\), which is the yield of clean coal by re - washing. Substitute \(a_2\) into the cumulative float curve (β curve) of floaters, so as to obtain the corresponding abscissa \(a_3\), which is the ash content of clean coal by re - washing.

[0120] 7. Determine the flotation process parameters: Input the boundary ash content λ, check the λ curve of the actual floatability curve, so as to obtain the corresponding ordinate \(b_2\), which is the yield of clean coal by flotation. Substitute \(b_2\) into the β curve to obtain the corresponding abscissa \(b_3\), which is the ash content of clean coal by flotation.

[0121] 8. Determine the product structure: Calculate the effects of each separation operation according to the principle of equal boundary ash content, calculate the total ash content of clean coal \(A_{calculated}\), so as to set the corresponding production parameters.

[0122] Obtain the total ash content of clean coal \(A_{calculated}\) according to the weighted average ash content of the ash content of clean coal by re - washing and the ash content of clean coal by flotation.

[0123] According to the principle of equal boundary ash content, the yield of clean coal obtained when the difference between the total ash content of clean coal and the target ash content of clean coal is less than the set error standard value is higher than the yield of clean coal obtained when the difference between the total ash content of clean coal and the target ash content of clean coal is greater than the set error standard value.

[0124] Set the error standard value ε, step size α and target ash content of clean coal \(A_{target}\).

[0125] If \(|A_{calculated}-A_{target}|>\varepsilon\):

[0126] When \(A_{calculated}>A_{target}\), \(A_{calculated}=A_{calculated}-\alpha\); when \(A_{calculated}<A_{target}\), \(A_{calculated}=A_{calculated}+\alpha\).

[0127] If \(|A_{calculated}-A_{target}|<\varepsilon\), the total ash content of clean coal \(A_{calculated}\) meets the requirements.

[0128] On the density curve within the actual washability curve of gravity separation, take the total ash content of clean coal as the ordinate value, and obtain the corresponding abscissa value as the separation density of the gravity separation link.

[0129] On the cumulative yield of floaters - number of distribution release test curves within the actual floatability curve, take the total ash content of clean coal as the ordinate value, and obtain the corresponding abscissa value as the optimal flotation stage number of the flotation link.

[0130] The present invention designs a calculation template for maximizing the yield of clean coal. When the properties of raw coal or the required product structure of users change, only the corresponding raw coal ash content or boundary ash content setting value in the template needs to be changed, and the template can output the theoretical ash content and product structure of each link, as well as the total yield and ash content of clean coal, and transmit its calculation results to the control module to achieve precise control of the product structure.

[0131] The clean coal yield maximization system based on big data includes a knowledge base, a database, an inference engine, an interpreter and a human-computer interaction interface. The knowledge base is used to store a database of professional knowledge. The database is used to record and store the input data, experimental data and calculation, reasoning and process data of experts and technicians in the expert system; the inference engine matches the appropriate output value according to the input value through the preset reasoning logic; the interpreter is used to modify or explain the output value of the inference engine; the human-computer interaction interface is used to establish a communication connection with the user interface to transmit instructions, display data and manually control parameter input. The database also includes empirical data, and the processing of the empirical data includes: storing original floating and sinking data, distribution release data and screening data under the corresponding coal quality and the proportion of gravity separation and flotation; when the equipment fails or the indicators are abnormal, the adjustment parameters manually input by the technicians and their corresponding requirements for gravity separation and flotation clean coal will be divided.

[0132] The present invention also has the following beneficial effects:

[0133] The present invention adopts a calculation method for maximizing the clean coal yield based on big data, which replaces the cumbersome series of experimental steps of traditional raw coal medium analysis. By predicting the properties of raw coal through big data, it can not only accurately and quickly determine the raw coal washing ratio, but also use the equal boundary ash principle to calculate the theoretical clean coal ash content of each link to achieve the maximum clean coal yield under the condition of required clean coal ash content. It can not only reduce the labor intensity of operators' laboratory tests, but also eliminate the interference caused by human factors, improve the efficiency and scientificity of formulating washing plans, and significantly improve the automation and intelligent level of the entire coal preparation process, ensuring that the coal preparation process can achieve the maximum clean coal yield, that is, the maximum economic benefit.

[0134] An embodiment of the present invention provides a system for determining clean coal yield, comprising:

[0135] The ash content prediction module is used to obtain the float-sink table containing the float-sink table yield and the historical gravity separation feed ash content in the case of gravity separation, and the step-by-step release data table containing the step-by-step release data table yield and the historical flotation feed ash content in the case of flotation.

[0136] The yield correction module is used to use the raw coal ash content and raw coal screening data to train the neural network model to predict the gravity separation feed ash content and the flotation feed ash content, and obtain the ash content prediction model; input the raw coal ash content to be tested into the ash content prediction model to obtain the gravity separation feed ash content and the flotation feed ash content; according to the difference between the gravity separation feed ash content and the historical gravity separation feed ash content, the floatation and sinking table yield is corrected and the floatation and sinking table is updated; according to the difference between the flotation feed ash content and the historical flotation feed ash content, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated floatation and sinking table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted.

[0137] The yield improvement module is used to obtain the gravity separation clean coal ash content and the flotation clean coal ash content in the actual selectivity curve and the actual floatability curve of gravity separation according to the boundary ash content used to distinguish the clean coal; according to the gravity separation clean coal ash content and the flotation clean coal ash content, the total clean coal ash content is obtained using the equal boundary ash principle, and the corresponding production parameters are set according to the total clean coal ash content to determine the clean coal yield.

[0138] A specific embodiment is as follows:

[0139] like Figure 1 The process of calculating the maximum clean coal yield of a coal preparation plant is shown below:

[0140] Among them, the calculation template for maximizing the clean coal yield is used: when the raw coal properties or the product structure required by the user change, only the corresponding raw coal ash content or boundary ash content setting value in the template needs to be changed, and the template can output the theoretical ash content and product structure of each link, as well as the total clean coal yield and ash content, and pass its calculation results to the control module to achieve precise control of the product structure. Specifically, the clean coal yield maximization system based on big data includes a knowledge base, a database, an inference engine, an interpreter and a human-computer interaction interface. The knowledge base is used to store a database of professional field knowledge, and the database is used to record and store the input data, experimental data and calculation, reasoning and process data of experts and technicians in the expert system; the inference engine matches the appropriate output value according to the input value through the preset reasoning logic; the interpreter is used to modify or explain the output value of the inference engine; the human-computer interaction interface is used to establish a communication connection with the user interface to transmit instructions, display data and input manual control parameters.

[0141] In this example, Matlab programming is used to implement the PB neural network to train the historical block float data and predict the floating and sinking composition of raw coal. Python programming is used to define the distribution rate function and output the distribution rate of each density. Access database is used as the comprehensive database of the expert system. The collected knowledge is stored in the database through sensor collection and manual entry. Based on the analysis of the collected ash data and the information provided by the on-site staff, a knowledge base of the working status of the coal flotation process is designed, and an expert system database based on historical fast floatation data, raw coal ash content, and clean coal ash content of each product is established.

[0142] In addition, the database also includes experience data, the processing of which includes: storing manual control parameters and their corresponding gravity flotation selection ratio, gravity clean coal ash content, and flotation clean coal ash content; when equipment fails or indicators are abnormal, the adjustment parameters manually input by technicians and their corresponding gravity flotation selection ratio, gravity clean coal ash content, and flotation clean coal ash content. The detected gravity flotation selection ratio, gravity clean coal ash content, and flotation clean coal ash content are sent to the database, and these data will be stored in the Access database of the industrial computer. At the same time, the database also includes rule data entered by flotation experts, technicians, and experienced workers.

[0143] The present invention comprises the following steps:

[0144] S1. The historical floating and sinking data of gravity separation feed and the floatability data (distribution release data) of flotation feed are collected through the coal preparation plant centralized control system.

[0145] S2. From the raw coal ash content and raw coal screening results, it can be seen that when the raw coal ash content changes, the yield of -1.5 density level and +1.8 density level has the largest change relative to the others, and the yield of -1.5 density level and +1.8 density level has the largest fluctuation relative to the baseline. Therefore, the change of raw coal ash content has the greatest impact on the yield of -1.5 density level and +1.8 density level. In the actual production process on site, sorting parameters such as sorting density are rarely changed. Therefore, when the quality of raw coal changes, under the condition that the sorting parameters do not change much, the change in the quality of raw coal is also reflected in the change in ash content of gravity separation products, clean coal, medium coal and gangue, that is, the yield of raw coal at -1.5 density level and the yield of +1.8 density level also have a certain influence relationship with the ash content of gravity separation products, clean coal, medium coal and gangue. Therefore, the ash content of raw coal, the ash content of gravity separation products, clean coal, medium coal and gangue can be used as independent variables, and the yield of raw coal at -1.5 density level and the yield of +1.8 density level can be used as dependent variables to train the model through BP neural network to determine the density composition of the coal, thereby predicting the ash content of gravity separation feed and flotation feed.

[0146] Because the values ​​of the four input parameters vary greatly, in order to avoid the difference in values ​​affecting the effect of the model, the data is scaled. Scaling includes data normalization and data standardization. This paper mainly uses data normalization to process data, mapping each sample data to between [-1, 1]. The formula is:

[0147]

[0148] where y max and min For parameters, the default values ​​are -1 and 1, x min is the minimum value of the original data to be processed, x max is the maximum value of the original data to be processed, and a is the original data after normalization. When outputting data, the data is denormalized.

[0149] S3. Correction of historical floating and sinking data: input the new reselected feed ash content - calculate the difference between the feed ash content and the original table ash content Δ weight - use Δ weight to correct the yield of the original floating and sinking table, and generate a new floating and sinking table based on the corrected yield and ash content. Figure 2 It is the corrected floating and sinking data table.

[0150] S4. Plotting of actual selectivity curve: Calculate the corresponding distribution rates of the five values ​​of sorting density of 1.3, 1.4, 1.5, 1.6 and 1.7, calculate the data table of actual selectivity curve of gravity sorting, fit and plot the actual selectivity curve of gravity sorting, Figure 3 It is the actual selectivity curve of gravity sorting drawn after fitting.

[0151] S5. Correction of historical step-by-step release data: Input the new flotation feed ash content, calculate the difference Δfloat between the feed ash content and the total ash content in the original step-by-step release data table. Use Δfloat to correct the yield of the original step-by-step release data table, generate a new distribution release data table based on the corrected yield and ash content, fit and draw the actual floatability curve, Figure 4 and Figure 5 They are respectively the corrected step-by-step release data table and the actual floatability curve graph drawn after fitting.

[0152] S6. Determine the gravity separation process parameters: according to the target ash content A, set the boundary ash content λ, check the λ curve of the actual selectivity curve of gravity separation, and obtain its corresponding ordinate a2, which is the gravity separation clean coal yield. Substitute a2 into the floating matter accumulation β curve to obtain the corresponding abscissa a3, which is the gravity separation clean coal ash content; substitute a2 into the density δ curve to obtain the corresponding upper abscissa a4, which is the gravity separation theoretical separation density.

[0153] S7. Determine flotation process parameters: input the boundary ash content λ, check the λ curve of the actual floatability curve, and obtain its corresponding ordinate b2, which is the flotation clean coal yield. Substitute b2 into the β curve to obtain the corresponding abscissa b3, which is the flotation clean coal ash content; substitute b2 into the v curve to obtain the corresponding abscissa b4, which is the optimal flotation number.

[0154] S8. Determine product structure: Calculate the effect of each sorting operation according to the equal boundary ash principle, calculate the total clean coal ash content A, and set corresponding production parameters at the same time; set a4 as the theoretical sorting density of gravity separation and b4 as the optimal flotation stage number.

[0155] The basic theory of the principle of minimum clean coal ash content and maximum clean coal yield is that, taking into account the recovery rate of combustibles, when two or more types of coal are sorted separately, in order to minimize the total clean coal ash content or maximize the total clean coal yield, the principle of equal basic ash content must be followed.

[0156] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for determining clean coal yield, characterized in that: The following steps are involved: Obtain a float-sink table containing the float-sink table yield and the historical gravity separation feed ash content in the case of gravity separation, and a step-by-step release data table containing the step-by-step release data table yield and the historical flotation feed ash content in the case of flotation; the gravity separation feed ash content indicates the content of the non-combustible components obtained after the mineral particles in the raw coal are separated by gravity separation, and the flotation feed ash content indicates the content of the non-combustible components obtained after the mineral particles in the raw coal are separated by flotation; The raw coal ash content is used to train the neural network model to predict the ash content of the gravity separation feed and the flotation feed, and an ash content prediction model is obtained; the raw coal ash content to be tested is input into the ash content prediction model to obtain the ash content of the gravity separation feed and the flotation feed; According to the difference between the ash content of the gravity separation feed and the historical gravity separation feed, the float-sink table yield is corrected and the float-sink table is updated; according to the difference between the ash content of the flotation feed and the historical flotation feed, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated float-sink table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted; According to the boundary ash content used to distinguish clean coal, the gravity separation clean coal ash content and flotation clean coal ash content are obtained in the actual selectivity curve and actual floatability curve of gravity separation respectively; according to the gravity separation clean coal ash content and the flotation clean coal ash content, the total clean coal ash content is obtained using the equal boundary ash content principle, and the corresponding production parameters are set according to the total clean coal ash content to determine the clean coal yield.

2. The method for determining the clean coal yield according to claim 1, characterized in that: The specific steps of correcting the yield of the float-sink table include: The difference Δweight between the ash content of the gravity separation feed and the ash content of the historical gravity separation feed is obtained by the formula: Δ reentry = λ reentry - λ reentry Wherein: λre-entry is the ash content of the gravity feed; λre-original is the ash content of the historical gravity feed, and the historical gravity feed ash content is specifically: the subtotal ash in the historical gravity feed float and sink table; Get density greater than +1.8g / cm 3 The density level of ash Ad +1.8 , the formula is; Where Adn means the nth one is greater than +1.8g / cm 3 The density level of ash; γn means the nth density level greater than +1.8g / cm 3 The yield of density grade; n means greater than +1.8g / cm 3 The number of density levels; Get density less than -1.8g / cm 3 The density level of ash Ad -1.8 , the formula is: Among them, Adm means the mth one is less than -1.8g / cm 3 The density level of ash; γm means the mth density level is less than -1.8g / cm 3 The yield of density grade; m means less than -1.8g / cm 3 The number of density levels; Get the calibration value X, the formula is: X = (100 * Δweight) / (Ad +1.8 -Ad -1.8 ) Correction of the yield of the float-sink table; If the density level is +1.8g / cm 3 : γ correction = γ + X If the density level is -1.8g / cm 3 : Here, γ correction represents the yield of the corresponding density level after correction, and γ represents the yield of the corresponding density level before correction.

3. The method for determining the clean coal yield according to claim 1, characterized in that: The specific steps of correcting the yield of the distribution release data table include: The difference Δfloat between the ash content of the flotation feed and the ash content of the historical flotation feed is obtained. The formula is: Δ float = λ float in - λ float out Among them, λ float in is the flotation feed ash; λ float original is the historical flotation feed ash, and the historical flotation feed ash is specifically: the weighted ash of each clean tailing coal in the historical distribution release data table; Get distribution release of the first tail coal ash Ad 轻 , the formula is; Among them, Adtailm represents the ash content of the mth tail coal released by distribution; γtailm represents the yield of the mth tail coal released by distribution; m represents the number of flotation in the distribution release test; Adfine represents the ash content of the clean coal released by distribution; γfine represents the yield of the clean coal released by distribution; γtail1 represents the yield of the first tail coal released by distribution; Get the calibration value X, the formula is: X = (100 * Δbuoyancy) / (Ad 尾1 -Ad 轻 ) Among them, Ad 尾1 Indicates the distribution and release of the ash content of the first tail coal; The distribution release table yield was corrected; For distribution and release of first flotation tailings: γ correction = γ tail 1 + X For the distribution release of the mth flotation tailings (m>=2): γ correction = γ tail m-γ tail m*X / (100-γ tail 1) For distribution release of clean coal: γ correction = γ essence - γ essence * X / (100-γ tail 1) Among them, γ correction represents the yield of clean (tail) coal after correction for the corresponding number of flotation times.

4. The method for determining the clean coal yield according to claim 1, characterized in that: The steps for obtaining the ash content of the heavy-medium clean coal from the actual washability curve of gravity separation are as follows: Set the cut-off ash for distinguishing clean coal according to the target ash content required for clean coal. On the ash-characteristic curve within the actual washability curve of gravity separation, take the cut-off ash as the abscissa value and obtain the corresponding ordinate value as the yield of heavy-medium clean coal; the actual washability curve of gravity separation includes: ash-characteristic curve λ, cumulative float curve β, and density curve δ. On the cumulative float curve within the actual washability curve of gravity separation, take the yield of heavy-medium clean coal as the ordinate value and obtain the corresponding abscissa value as the ash content of the heavy-medium clean coal.

5. The method for determining the clean coal yield according to claim 1, characterized in that: The steps for obtaining the ash content of the flotation clean coal from the actual floatability curve are as follows: On the ash-characteristic curve within the actual floatability curve, take the cut-off ash as the abscissa value and obtain the corresponding ordinate value as the yield of flotation clean coal; the actual floatability curve includes: ash-characteristic curve λ, cumulative float curve β, cumulative float yield - distribution release test times curve v, and cumulative sink curve n. On the cumulative float curve within the actual floatability curve, take the yield of flotation clean coal as the ordinate value and obtain the corresponding abscissa value as the ash content of the flotation clean coal.

6. A method for determining clean coal yield according to claim 1, characterized in that: The steps for obtaining the total clean coal ash using the equal cut-off ash principle are as follows: Obtain the total clean coal ash content A_calculated based on the weighted average ash content of the heavy-medium clean coal ash and the flotation clean coal ash. According to the equal cut-off ash principle, when the difference between the total clean coal ash and the target clean coal ash is less than the set error standard value, the yield of clean coal obtained is higher than that when the difference between the total clean coal ash and the target clean coal ash is greater than the set error standard value. Set the error standard value ε, step size α, and target clean coal ash A_target. If |A_calculated - A_target| > ε: When A_calculated > A_target, A_calculated = A_calculated - α; when A_calculated < A_target, A_calculated = A_calculated + α. If |A_calculated - A_target| < ε, the total clean coal ash content A_calculated meets the requirements.

7. A method for determining clean coal yield according to claim 1, characterized in that: The steps for setting the corresponding production parameters based on the total clean coal ash are as follows: On the density curve within the actual washability curve of gravity separation, take the total clean coal ash as the ordinate value and obtain the corresponding abscissa value as the separation density of the gravity separation stage. On the cumulative float yield - distribution release test times curve within the actual floatability curve, take the total clean coal ash as the ordinate value and obtain the corresponding abscissa value as the optimal number of flotation stages in the flotation stage.

8. A system for determining clean coal yield, characterized in that: Include: An ash prediction module for obtaining a float-sink table containing the yield of the float-sink table and the historical heavy-medium feed ash content in the case of gravity separation, and a step-by-step release data table containing the yield of the step-by-step release data table and the historical flotation feed ash content in the case of flotation. The yield correction module is used to use the raw coal ash content to train the neural network model to predict the gravity separation feed ash content and the flotation feed ash content to obtain the ash content prediction model; input the raw coal ash content to be detected into the ash content prediction model to obtain the gravity separation feed ash content and the flotation feed ash content; according to the difference between the gravity separation feed ash content and the historical gravity separation feed ash content, the float-sink table yield is corrected and the float-sink table is updated; according to the difference between the flotation feed ash content and the historical flotation feed ash content, the distribution release data table yield is corrected and the distribution release data table is updated; according to the updated float-sink table, the actual selectivity curve of gravity separation is fitted; according to the updated distribution release data table, the actual floatability curve is fitted; The yield improvement module is used to obtain the gravity separation clean coal ash content and the flotation clean coal ash content in the actual selectivity curve and the actual floatability curve of gravity separation according to the boundary ash content used to distinguish the clean coal; according to the gravity separation clean coal ash content and the flotation clean coal ash content, the total clean coal ash content is obtained using the equal boundary ash principle, and the corresponding production parameters are set according to the total clean coal ash content to determine the clean coal yield.

Images