Coal blending planning method based on geometric constraint of coal rock reflectivity distribution curve

Abstract

The invention provides a coal blending planning method based on coal rock reflectivity distribution curve geometric constraint, and relates to the technical field of coking coal blending. The coal blending method is characterized in that constraint conditions are constructed by using coal rock reflectivity distribution curves in historical production data, nonlinear optimization solution is carried out, and the coal blending method comprises the following steps: extracting the coal rock reflectivity distribution curves of historical successful coal blending schemes, forming historical successful curve clusters, and further constructing safe coal rock distribution envelope bands; converting the constraint condition into an additional constraint condition in the planning model, integrating the constraint condition for the coal blending quality index and the constraint condition for the proportion index, constructing the planning model by taking the lowest coal blending cost as an objective function, and solving to obtain the optimal coal blending proportion. According to the method, accurate constraint conditions are constructed through the full-waveform coal rock reflectivity distribution curve, the stability and prediction accuracy of coke quality are remarkably improved, the coal blending cost per ton of coke is minimized, and the research and development period of new coal introduction or scheme adjustment is shortened.

Description

Technical Field

[0001] This invention relates to the field of coking coal blending technology, and in particular to a coal blending planning method based on the geometric constraints of coal and rock reflectance distribution curves. Background Technology

[0002] Existing automatic coal blending systems are mainly based on linear programming, with the goal of satisfying chemical composition (ash content Ad, sulfur content Std, volatile matter Vdaf) and specific petrographic indicators. The lowest cost is achieved under the premise of [missing information]. Existing technologies have the following shortcomings:

[0003] (1) Information loss: traditional It is just an average value and cannot reflect the "mixed coal structure" of the blended coal (such as whether bimodal or concave depths exist). Blended coals with the same average reflectance may have drastically different distributions of active and inert components, leading to fluctuations in coke quality (CSR / CRI).

[0004] Specifically, this manifests in the quality dimension: the "average trap" caused by existing single-index models and the problem of coke quality fluctuations, where current technologies rely solely on the average reflectance of the vitrinite group. Linear programming of statistical indicators fails to identify and differentiate technical defects in coal blending structures with the same average value but drastically different microscopic distribution patterns (e.g., the inability to distinguish between "normal distribution" and "bimodal distribution" or inferior blended coal with "intermediate gaps"). This defect leads to distortion in the prediction model, resulting in uncontrollable fluctuations in the cold strength (M40 / M10) and hot properties (CSR / CRI) of the produced coke.

[0005] (2) Nonlinear relationship: There is often a nonlinear relationship between coke quality and the proportion of single coal types. The simple linear superposition model has low prediction accuracy.

[0006] Specifically, this manifests in the cost dimension: Insufficient model prediction accuracy leads to inflated coal blending costs. Existing linear weighted models cannot accurately reflect the complex nonlinear interactions in the coking process, forcing companies to reserve excessive "quality safety margins" in their coal blending schemes in order to ensure the quality of coke. This means excessively adding high-priced, high-quality prime coking coal or fat coal, resulting in persistently high raw coal costs.

[0007] (3) Experience is difficult to reuse: The successful experience of experienced workers is often hidden in coal and rock charts and is difficult to quantify into mathematical models.

[0008] Specifically, this is reflected in the dimensions of efficiency and energy saving: the new solution has a long development cycle and the small coke oven experiment has high energy consumption. The traditional coal blending method mainly relies on manual experience trial blending and must be physically verified through high-frequency small coke oven experiments (40kg or larger scale), which results in a long coal blending solution adjustment cycle (usually 3-5 days), high experimental energy consumption, and huge human and time costs. Summary of the Invention

[0009] To address the shortcomings of existing technologies, this invention proposes a coal blending planning method based on the geometric constraints of coal-rock reflectance distribution curves. Specifically, it transforms the microstructural characteristics (coal-rock curves) of historically successful coal blending into an optimization method with mathematical constraints. The technical problems to be solved include:

[0010] (1) By introducing the full-waveform coal and rock reflectivity distribution curve as a geometric constraint, the problem of information loss is solved, and the substantial similarity between the new coal blending scheme and the historical high-quality scheme is ensured from the microscopic rock facies structure level, thereby significantly improving the stability of coke quality and the prediction accuracy.

[0011] (2) Utilize the successful experience of historical big data to construct a precise nonlinear “safety envelope”. Under the premise of ensuring that the quality of coke is not lower than the target value (such as CSR>65), accurately explore the blending potential of low-priced coal types (such as weakly caking coal and gas coal) and eliminate unnecessary quality excess, thereby minimizing the cost of coal blending per ton of coke at a quantitative level.

[0012] (3) By digitally reusing the coal and rock fingerprint characteristics of historically successful coal blending, more than 90% of schemes that theoretically meet the conventional indicators but have unreasonable microstructures are eliminated in the calculation stage, which greatly reduces the number of physical experiments required, shortens the R&D cycle of introducing new coal types or adjusting schemes, and indirectly reduces energy consumption and carbon emissions in the R&D process.

[0013] In a first aspect, the present invention provides a coal blending planning method based on the geometric constraints of the coal and rock reflectivity distribution curve, comprising the following steps:

[0014] S1. Input the quality indicators of the target coke;

[0015] S2. Screen out all historical coal blending schemes whose actual production results meet the quality indicators, extract the corresponding coal-rock reflectance distribution curves of the blended coal, and form a cluster of historical success curves.

[0016] S3. Based on the historical success curve cluster, construct the safe coal and rock distribution envelope and transform it into an additional constraint in the planning problem;

[0017] S4. Construct a planning model with the objective function of minimizing the cost of blended coal. The constraints include constraints on the quality indicators of blended coal, constraints on the proportion indicators, and additional constraints obtained in step S3.

[0018] S5. Solve the planning model to obtain the optimal coal blending ratio.

[0019] As a further improvement of the present invention, in the constraint conditions of step S3, the constraint function is the coal rock reflectance distribution curve of the blended coal.

[0020] As a further improvement of the present invention, the constraint construction method in step S3 is: envelope constraint and / or similarity distance constraint, wherein,

[0021] Envelope constraints are constructed by calculating the statistical upper and lower boundaries of historical success curve families;

[0022] The similarity distance constraint is calculated by determining the center curve of the historical success curve cluster, and the constraint is constructed by limiting the distance between the coal and rock reflectance distribution curve and the center curve.

[0023] As a further improvement of the present invention, the specific method for constructing the envelope constraint is as follows:

[0024] ;

[0025] in, Let j be the mass percentage of the j-th type of coal in the blended coal. Let be the vector of the reflectance distribution curve of the j-th type of single coal. , express For the portion within the k-th reflectivity interval, K is the total number of intervals;

[0026] , This represents the statistical lower and upper boundaries of the historical successful curve cluster within the k-th reflectance interval of the coal and rock reflectance distribution, which includes boundary scaling through a relaxation factor.

[0027] As a further improvement of the present invention

[0028] The center curve is a mean curve;

[0029] The distance is the square of the Euclidean distance.

[0030] As a further improvement of the present invention, the specific method for constructing the similarity distance constraint is as follows:

[0031] ;

[0032] in, Let j be the mass percentage of the j-th type of coal in the blended coal. Let be the vector of the reflectance distribution curve of the j-th type of single coal. , express For the portion within the k-th reflectivity interval, K is the total number of intervals; The mean curve of the historical success curve cluster The kth component, The maximum allowable difference threshold.

[0033] As a further improvement of the present invention, in step S4,

[0034] The cost of the blended coal is calculated using the price and proportion of each type of coal.

[0035] The coal blending quality index is the content ratio of each chemical component in the blended coal, which is obtained by weighted calculation of each individual coal in the blended coal;

[0036] The constraint on the ratio index is: the constraint on the maximum allowable mass percentage of a single type of coal under inventory and process conditions.

[0037] As a further improvement of the present invention, the solution method for step S5 includes:

[0038] For the case where only envelope constraints are used, the simplex method is employed for solution;

[0039] For cases where similarity distance constraints are used, or where a nonlinear coke quality prediction model is further superimposed as an additional constraint in step S3, a sequential quadratic programming or interior point method solver is used to solve the problem.

[0040] In a second aspect, the present invention provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.

[0041] Thirdly, the present invention provides a computer program product that, when executed by a processor, implements the steps of the method described in the first aspect.

[0042] Compared with existing technologies, this invention constructs precise nonlinear constraints by using full-waveform coal and rock reflectivity distribution curves, which significantly improves the stability and prediction accuracy of coke quality, minimizes the cost of coal blending per ton of coke, and shortens the R&D cycle for introducing new coal types or adjusting schemes. Attached Figure Description

[0043] Figure 1 This is a flowchart of a coal blending planning method based on the geometric constraints of coal and rock reflectivity distribution curves, as disclosed in this invention. Detailed Implementation

[0044] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Steps S1, S2… in the described embodiments of the present invention do not limit the scope of execution of the present invention; the various models, simulation environments, and software described in the present invention are not considered as the only limiting methods of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0045] In this invention, computer device / equipment / system refers to a related entity applied to a computer, such as hardware, a combination of hardware and software, software, or software in execution. More specifically, for example, software includes, but is not limited to, a process running on a processor, a processor, an object, executable software, an execution thread, a program, and / or a computer. Furthermore, an application program or script running on a server, and the server itself, can also be software. One or more software programs may be in an execution process and / or thread, and the software may be localized on one computer and / or distributed across two or more computers, and may be run on various computer-readable media.

[0046] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0047] In a first aspect, the present invention provides an embodiment of a coal blending planning method based on the geometric constraints of coal and rock reflectivity distribution curves, such as... Figure 1 As shown, the specific process can be as follows:

[0048] S1. Demand Input: The user sets the target quality requirements for the coke.

[0049] In one embodiment of the present invention, the target coke quality index is:

[0050] CSR≥65, CRI≤25, M40>85.

[0051] S2, Historical Optimization:

[0052] S21. Screening and feature extraction of historical successful samples: Traverse the historical database and screen out all historical coal blending schemes whose actual production results meet the quality requirements set in step S1.

[0053] Specifically, define a set of success histories. :

[0054] ;

[0055] Suppose that the set contains M historical schemes, where, The actual coke quality result corresponding to the i-th scheme in the historical database. : User-defined target coke quality requirement vector : Represents the index of coal blending schemes in the historical database.

[0056] S22. Extract the corresponding set of successful historical coal blending curves.

[0057] Extract the coal-rock reflectance distribution curves corresponding to the schemes in the successful historical set, i.e., the historical successful curve cluster:

[0058] ;

[0059] in It is a collection of successful histories The actual coal blending curve of the i-th successful scheme recorded in the data.

[0060] S3. Feature Extraction and Constraint Construction:

[0061] Statistical analysis was performed on multiple historical success curves to construct a safe coal and rock distribution envelope.

[0062] This envelope is transformed into an additional constraint in the nonlinear optimization model.

[0063] This invention provides two implementation methods for constructing constraints, which can be used either one or simultaneously:

[0064] Implementation method 1: Envelope constraint, i.e., linear constraint;

[0065] (1) For each reflectance interval k, calculate the statistical upper and lower boundaries of the historical success curve cluster:

[0066] ;

[0067] ;

[0068] in, : Represents an index of coal blending schemes in the historical database. This represents the reflectance interval index in the coal and rock reflectance distribution map. For example, k=1 represents the interval [0.50, 0.55), and K is usually taken as 50. is the relaxation factor for the k-th interval, used to improve the feasibility of the model.

[0069] (2) The corresponding constraint expression is:

[0070] ;

[0071] in, This represents the index of the reflectance interval in the coal and rock reflectance distribution map. , These represent the statistical lower and upper boundaries of the historical success curve cluster, respectively. , indicating the index of the single type of coal to be selected; Let be the coal-lithology curve vector for a single type of coal, and represent the vitrinite reflectance frequency distribution vector for the j-th type of single coal. ,in This represents the content of this coal type in the k-th reflectivity interval, and satisfies... That is, the total vitrinite content is 1.

[0072] Implementation Method 2: Similarity distance constraint, i.e., nonlinear constraint;

[0073] (1) Calculate the center curve of the historical success curve cluster;

[0074] Preferably, the mean curve is used as the center curve:

[0075] ;

[0076] in, It is a cluster of historical success curves, specifically a collection of successful historical events. It includes M historical schemes.

[0077] (2) Define the distance metric function ;

[0078] Preferably, the squared Euclidean distance (L2 Norm) is used:

[0079] ;

[0080] in, Mean curve The k-th component represents the portion of the mean curve within the k-th reflectance interval. C(x) is the coal-rock curve vector of the blended coal, which is the predicted reflectance distribution vector of the blended coal determined by the decision variable x.

[0081] Specifically, C(x) is obtained based on the weighted superposition principle:

[0082] ;

[0083] in, Let j be the mass percentage of the j-th type of coal in the new coal blending scheme. Let be the vitrinite reflectance frequency distribution vector of the j-th type of coal.

[0084] (3) The corresponding constraint expression is:

[0085] ;

[0086] in, Let C(x) be the k-th component. Mean curve The kth component, The maximum allowable difference threshold.

[0087] S4. Construct a nonlinear programming model:

[0088] A nonlinear programming model is constructed with the objective function of minimizing coal blending cost and the constraints of conventional coal quality indicators and the aforementioned coal-rock curve characteristics.

[0089] The nonlinear programming model is constructed as follows:

[0090] Objective function: Minimize coal blending cost.

[0091] ;

[0092] in, Let j be the mass percentage of the j-th type of coal in the new coal blending scheme. Let be the price of the j-th type of coal.

[0093] Constraints:

[0094] (1) Basic quality constraints, for example, constraints on ash and sulfur content are as follows:

[0095] ;

[0096] ;

[0097] in, These represent the proportions of ash and sulfur in the j-th type of coal, respectively. These are the upper limits for the proportions of ash and sulfur, respectively;

[0098] (2) Proportioning constraints:

[0099] ;

[0100] ;

[0101] Wherein, is the maximum mass percentage of the j-th type of coal under inventory and process conditions.

[0102] (3) Coal and rock curve morphology constraints:

[0103] Preferably, the envelope constraint of Implementation Method 1 or the similarity distance constraint of Implementation Method 2 in step S3 can be used, specifically the constraint expression therein.

[0104] S5. Solve for the optimal proportions of various types of single coal.

[0105] Solution methods include:

[0106] (1) If only the envelope constraint is used, the model is a linear programming (LP) problem, which is solved by the simplex method.

[0107] (2) If a similarity distance constraint is adopted, or a nonlinear coke quality prediction model is introduced as an additional constraint, then the model is a nonlinear programming (NLP) problem, which is solved by a sequential quadratic programming (SQP) or interior point method solver.

[0108] In a second aspect, the present invention provides an embodiment of a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the first aspect.

[0109] Thirdly, the present invention provides a computer program product embodiment, which, when executed by a processor, implements the steps of the method described in the first aspect.

Claims

1. A coal blending planning method based on the geometric constraints of coal and rock reflectivity distribution curves, characterized in that, Includes the following steps: S1. Input the quality indicators of the target coke; S2. Screen out all historical coal blending schemes whose actual production results meet the quality indicators, extract the corresponding coal-rock reflectance distribution curves of the blended coal, and form a cluster of historical success curves. S3. Based on the historical success curve cluster, construct the safe coal and rock distribution envelope and transform it into an additional constraint in the planning problem; S4. Construct a planning model with the objective function of minimizing the cost of blended coal. The constraints include constraints on the quality indicators of blended coal, constraints on the proportion indicators, and additional constraints obtained in step S3. S5. Solve the planning model to obtain the optimal coal blending ratio.

2. The method according to claim 1, characterized in that, In the constraints of step S3, the constraint function is the coal rock reflectance distribution curve of the blended coal.

3. The method according to claim 2, characterized in that, The constraint construction method in step S3 is: envelope constraint and / or similarity distance constraint, wherein, Envelope constraints are constructed by calculating the statistical upper and lower boundaries of historical success curve families; The similarity distance constraint is calculated by determining the center curve of the historical success curve cluster, and the constraint is constructed by limiting the distance between the coal and rock reflectance distribution curve and the center curve.

4. The method according to claim 3, characterized in that, The specific method for constructing the envelope constraint is as follows: ； in, Let j be the mass percentage of the j-th type of coal in the blended coal. Let be the vector of the reflectance distribution curve of the j-th type of single coal. , express For the portion within the k-th reflectivity interval, K is the total number of intervals; , This represents the statistical lower and upper boundaries of the historical successful curve cluster within the k-th reflectance interval of the coal and rock reflectance distribution, which includes boundary scaling through a relaxation factor.

5. The method according to claim 3, characterized in that, The center curve is a mean curve; The distance is the square of the Euclidean distance.

6. The method according to claim 5, characterized in that, The specific method for constructing the similarity distance constraint is as follows: ； in, Let j be the mass percentage of the j-th type of coal in the blended coal. Let be the vector of the reflectance distribution curve of the j-th type of single coal. , express For the portion within the k-th reflectivity interval, K is the total number of intervals; The mean curve of the historical success curve cluster The kth component, The maximum allowable difference threshold.

7. The method according to claim 1, characterized in that, In step S4 The cost of the blended coal is calculated using the price and proportion of each type of coal. The coal blending quality index is the content ratio of each chemical component in the blended coal, which is obtained by weighted calculation of each individual coal in the blended coal; The constraint on the ratio index is: the constraint on the maximum allowable mass percentage of a single type of coal under inventory and process conditions.

8. The method according to claim 3, characterized in that, The solution method for step S5 includes: For the case where only envelope constraints are used, the simplex method is employed for solution; For cases where similarity distance constraints are used, or where a nonlinear coke quality prediction model is further superimposed as an additional constraint in step S3, a sequential quadratic programming or interior point method solver is used to solve the problem.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-8.

10. A computer program product, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-8.

Images