A coking process energy consumption diagnosis method based on dynamic working condition matching and layered contribution degree analysis

Abstract

The present application relates to the metallurgical technical field, and particularly relates to a coking process energy consumption diagnosis method based on dynamic working condition matching and layered contribution degree analysis. The method constructs a rolling diagnosis period at the current diagnosis time, collects coking process production and energy data and establishes a production operation database, and constructs a unit coke comprehensive energy consumption model; the advanced benchmark working condition is obtained by screening the historical working condition samples meeting the preset constraints from the database, the comparable historical working condition is screened out by comparing the characteristic vectors, and the working condition with lower unit coke comprehensive energy consumption and stable operation is selected from the comparable historical working condition as the advanced benchmark working condition; the unit coke comprehensive energy consumption is decomposed into primary influencing factors, the influence value and contribution degree of each primary influencing factor are calculated, and the primary influencing factor diagnosis conclusion is obtained; the primary influencing factor with the absolute value of the contribution degree greater than the preset threshold value is selected, the corresponding secondary parameter value is obtained, the secondary parameter diagnosis parameter is calculated, and the diagnosis conclusion and adjustment suggestion are output.

Description

Technical Field

[0001] This invention relates to the field of metallurgical technology, and in particular to a method for diagnosing energy consumption in coking processes based on dynamic operating condition matching and stratified contribution analysis. Background Technology

[0002] Coking is a crucial upstream process in integrated steel production lines, responsible for coking coal input, carbonization, coke production, coke oven gas recovery, and the recovery and utilization of byproducts such as tar, crude benzene, and steam. The levels of gas consumption, electricity consumption, steam recovery, and byproduct recovery in the coking process directly impact the energy consumption per unit product and the energy balance of the enterprise. Therefore, conducting stable and accurate energy consumption assessments and exploring energy-saving potential in the coking process is of significant production management value and energy conservation and carbon reduction implications.

[0003] Currently, the unit energy consumption of the coking process is typically calculated using data from multiple systems, including production statistics, energy metering, and material settlement. The production statistics system provides output data such as coke quantity; the energy metering system provides consumption data for coking coal, non-coking coal, blast furnace gas, coke oven gas, electricity, steam, nitrogen, and water; and the material settlement system provides output or recovery data for tar, crude benzene, external supply of coke oven gas, and steam recovery. Existing statistical methods can obtain the comprehensive unit energy consumption of coke for a fixed statistical period, but they struggle to explain the specific reasons for increases or decreases in actual energy consumption as production conditions change.

[0004] Current energy consumption analyses in coking processes often compare indicators from fixed time periods, shift averages, monthly averages, or historical averages. However, because production load, gas calorific value, coal moisture content, carbonization time, vertical operating temperature, energy medium consumption, and waste energy recovery levels continuously change with operating conditions, fixed average comparisons struggle to distinguish between differences in operating conditions and operational deviations. Furthermore, it is difficult to decompose energy consumption differences into understandable impacts such as fuel consumption, energy conversion deficit, electricity consumption, and waste energy recovery.

[0005] Furthermore, existing industrial intelligent diagnostic solutions are mostly geared towards equipment failure, single parameter adjustments, or energy consumption prediction. Even when prediction results are provided, they typically lack a unified energy balance for energy input, energy recovery, and byproducts in the coking process. This makes it impossible to automatically match actual operating conditions with similar advanced benchmark conditions and decompose energy consumption differences. Although on-site personnel can see energy consumption deviations or predicted values, they cannot quickly obtain the contribution ranking of key influencing factors, the difference between actual and benchmark values, and the corresponding actionable parameters. This results in diagnostic conclusions that are difficult to interpret and recommendations that lack specificity. Summary of the Invention

[0006] To address the technical problems of existing statistical methods failing to explain the specific reasons for the increase or decrease in actual energy consumption as production conditions change, failing to distinguish between differences in operating conditions and operational deviations, failing to decompose energy consumption differences into understandable impact components, and failing to achieve automatic matching of actual operating conditions with similar advanced benchmark operating conditions and decomposition of energy consumption differences, this invention provides an energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis.

[0007] Therefore, the present invention provides the following technical solution: A method for diagnosing energy consumption in coking processes based on dynamic operating condition matching and hierarchical contribution analysis includes the following steps: S1. Based on the current diagnostic time t The end time is [time value], and the construction time is [duration value]. Δt The rolling diagnostic cycle collects historical production and energy data for the coking process within that cycle. S2. Based on the collected historical production and energy data, construct a production operation database for the coking process; S3. Based on the coking process production operation database, construct a comprehensive energy consumption model per unit of coke; S4. Extract historical operating condition samples from the coking process production operation database that have the same time window length as the current performance cycle and meet the preset constraints. Construct a historical operating condition feature vector containing preset index parameters based on each historical operating condition sample. Construct a current performance cycle operating condition feature vector containing the same preset index parameters based on the current performance operating condition. Compare the historical operating condition feature vector with the current performance cycle operating condition feature vector. Select historical operating conditions that meet the preset similarity conditions from the historical operating condition samples as comparable historical operating conditions. S5. Calculate the unit coke comprehensive energy consumption of each comparable historical working condition through the unit coke comprehensive energy consumption model, and sort them from low to high according to the unit coke comprehensive energy consumption. Select the comparable historical working condition with the lowest energy consumption value or within the preset quantile range and with stable operation as the advanced benchmark working condition. S6. According to the energy income and expenditure category, the unit coke comprehensive energy consumption under the current performance cycle operating condition and the advanced benchmark operating condition determined from the comparable historical operating conditions is decomposed into several first-level influencing factors. Based on the unit coke comprehensive energy consumption, the energy consumption component of the first-level influencing factors under the current performance cycle operating condition and the advanced benchmark operating condition is calculated respectively. The difference between the energy consumption components of the first-level factors under the current performance cycle operating condition and the advanced benchmark operating condition is determined as the first-level influence value of the first-level influencing factor. The total change in unit coke comprehensive energy consumption between the current actual operating condition and the advanced benchmark operating condition is calculated based on the energy balance relationship. The ratio of each primary influence value to the total change in unit coke comprehensive energy consumption is calculated to obtain the contribution of each primary influence factor. The factors are sorted from largest to smallest according to the absolute value of the first-level impact value, and the diagnosis of whether each first-level impact factor leads to an increase or decrease in energy consumption in the coking process is determined based on the sign of the first-level impact value. S7. Select the primary influencing factors whose absolute contribution value is greater than the preset threshold, obtain the values ​​of the secondary parameters of each primary influencing factor under the current performance cycle conditions and the advanced benchmark conditions determined from comparable historical conditions; calculate the diagnostic parameters of each secondary parameter based on these values, and output the diagnostic conclusions and adjustment suggestions for each secondary parameter.

[0008] Furthermore, the formula for calculating the rolling diagnostic cycle in step S1 is: [ t- Δ t,t ] The rolling diagnostic cycle follows a preset update step. δt Rolling updates, among which δt Less than or equal to Δ t .

[0009] Furthermore, the historical production and energy data of the coking process collected during the rolling diagnostic cycle are processed by unifying time windows, removing duplicate records, marking missing values, identifying outliers, and unifying units. For cumulative data, the increment within the rolling diagnostic cycle is calculated. For flow rate, temperature, and calorific value data, the cumulative value, average value, or characteristic value of the rolling diagnostic cycle is calculated based on the data attributes, and then converted into unit coke index with the coke production of the rolling diagnostic cycle as the denominator. Records of working conditions with zero coke production, missing metering, exceeding reasonable range, incomplete time windows, or violating preset material and energy constraints are marked as invalid.

[0010] Furthermore, in step S3, based on the coking process production operation database, a comprehensive energy consumption model per unit of coke is constructed, and the calculation formula is as follows:

[0011] In the formula: E coking The unit coke comprehensive energy consumption of the coking process during the rolling diagnostic cycle is expressed in kg / t. E coal , E' cog E Tar , E ber , E coke , E cog , E bfg , E dljz , E steam ,E ele The corresponding standard coal equivalents (kg) for washed clean coal, coke oven gas, tar, benzene, coke, coke oven gas and blast furnace gas, power medium, recovered steam, and self-generated electricity during the diagnostic period are: p coke The coke production during the diagnostic cycle is expressed in tons (t).

[0012] Furthermore, after obtaining the advanced benchmark operating conditions in step S5, the unit coke comprehensive energy consumption for the current performance period is calculated using the unit coke comprehensive energy consumption model. E actual Comprehensive energy consumption per unit of coke corresponding to advanced benchmark operating conditions E base According to Δ E = E actual - E base The difference in unit coke comprehensive energy consumption between the current performance period and the advanced benchmark operating condition is calculated; where Δ E If the value is greater than zero, it is determined that there is energy-saving potential in the current operating condition, and step S6 is continued; Δ E If the value is less than or equal to zero, it is determined that there is no energy-saving potential in the current operating condition, and the next step is stopped.

[0013] Furthermore, in step S6, the first i The first-level influence value of each first-level influencing factor I i The calculation formula is as follows:

[0014] Total change in comprehensive energy consumption per unit of coke ΔC The calculation formula is as follows:

[0015] The formula for calculating the contribution of primary influencing factors is as follows:

[0016] In the formula: or i For the first i The contribution of each primary influencing factor to the change in energy consumption in the coking process; C i,actual For the first i The energy consumption components of each primary influencing factor are uniformly converted according to the positive and negative directions in the unit coke comprehensive energy consumption model under the current actual performance cycle conditions. C i,base For the first iThe energy consumption components of each primary influencing factor are calculated using the same caliber under advanced benchmark conditions determined from comparable historical operating conditions. ΔC This represents the total change in energy consumption during the coking process; m The number of primary influencing factors; k The subscript is used for summation.

[0017] Furthermore, the diagnostic conclusions of each primary influencing factor output in step S6 include the primary factor energy consumption component under the current actual performance cycle operating conditions, the primary factor energy consumption component under the advanced benchmark operating conditions determined from comparable historical operating conditions, the primary relative change rate, the primary contribution, the ranking result, and the diagnostic conclusion that the primary influencing factor causes an increase or decrease in the energy consumption of the coking process. The formula for calculating the relative rate of change is as follows:

[0018] In the formula: r i For the first i The relative rate of change of each primary influencing factor is used to characterize the degree of deviation of the energy consumption component of the primary factor from the advanced benchmark under the current actual operating conditions. when When this happens, the first-order relative rate of change is not calculated, and a prompt indicating that the baseline value is zero is output. when I When the value is greater than 0, it is determined that this primary influencing factor leads to an increase in energy consumption in the coking process; when... I When <0, it is determined that this primary influencing factor leads to a reduction in energy consumption in the coking process; when I When the value is 0, it is determined that this primary influencing factor has no significant impact on the energy consumption changes in the coking process.

[0019] Furthermore, the diagnostic conclusions for each secondary parameter include the value of the secondary parameter in the current performance period, the value of the secondary parameter under the advanced benchmark condition determined in comparable historical conditions, the relative change rate of the secondary parameter, the secondary contribution, the ranking results based on the absolute value of the secondary contribution, and the diagnostic conclusion that the secondary parameter causes the corresponding primary influencing factor to increase or decrease. The formula for calculating the second-order relative rate of change is as follows:

[0020] In the formula: r j For the first j The relative rate of change of each second-order parameter; x j,actual For the first j The values ​​of each secondary parameter in the current performance period; x j,base For the firstj The values ​​of each secondary parameter under the advanced benchmark condition determined in comparable historical operating conditions; when x j,base When =0, the second-order relative rate of change is not calculated, and a prompt indicating that the baseline value is zero is output. The formula for calculating the sensitivity of each secondary parameter is as follows:

[0021]

[0022] In the formula: S j For the second-level parameters, the first j Sensitivity corresponding to each variable value; E j For the second-level parameters, the first j Energy consumption values ​​corresponding to each variable value; P j For the second-level parameters, the first j One variable value; E o This represents the energy consumption value corresponding to the basic variable value under the secondary parameters; P o The reference variable value under the second-level parameter; S average This represents the average sensitivity of the variable values ​​under the second-order parameter, i.e., the sensitivity coefficient. n This represents the total number of variables under the second-level parameters; Based on the secondary parameters and the comparable historical operating data corresponding to each secondary parameter, a multiple linear regression curve is used to establish the relationship between the secondary parameters and the corresponding primary influencing factors. The calculation formula is as follows:

[0023] In the formula: y , where is the target variable, representing the energy consumption impact value of the primary influencing factor corresponding to the secondary parameter; β 0 represents the intercept term; β 1. β 2、…、 β n These are the regression coefficients obtained by fitting a sample of comparable historical operating conditions. x 1. x 2、…、 x j These are the secondary parameters that were selected for modeling after sensitivity analysis; j The number of secondary parameters involved in the modeling; e This is the error term; The formula for calculating the secondary contribution is as follows:

[0024] In the formula, or j For the first j The contribution of each secondary parameter; Δ y j For the first j The change in energy consumption caused by the transformation of a secondary parameter from its value under the advanced benchmark condition determined in comparable historical operating conditions to its value under the current actual operating condition; Δ y j The sign indicates the direction of the effect of the secondary parameter on energy consumption, whether it increases or decreases.

[0025] Advantages and positive effects of the present invention: The system sets up a rolling diagnostic cycle to collect on-site production and energy data, adapting to the continuous production characteristics of the coking process. It can continuously conduct energy consumption analysis along with the production process, getting rid of the analysis lag problem caused by fixed statistical cycles and meeting the needs of routine energy consumption evaluation.

[0026] Comparable historical operating conditions are obtained through conditional filtering and feature vector similarity matching, which are used for energy consumption benchmarking analysis. This can distinguish between energy consumption changes caused by fluctuations in the operating conditions themselves and deviations in human operation, reduce the interference of inconsistent operating conditions on the analysis results, and improve the reliability of energy consumption comparison.

[0027] Instead of simply outputting overall energy consumption results, it can analyze the specific reasons for the increase or decrease in energy consumption, break down the total energy consumption deviation into various primary influencing factors, quantify the degree of influence of each part, and make the source of energy consumption fluctuations clearly visible.

[0028] Based on the analysis of primary influencing factors, the factors with significant contributions are further traced back to secondary parameters, achieving a step-by-step positioning from macro-level energy consumption differences to specific process parameters, grasping the main influencing objects, and avoiding meaningless calculations of all parameters.

[0029] The analysis results of the secondary parameters are combined to output diagnostic conclusions and adjustment suggestions, making it easy for on-site personnel to understand the analysis results and directly apply them to production optimization, providing practical support for the refined management and control of the coking process and energy conservation and carbon reduction. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1The present invention provides a general flowchart of a coking process energy consumption diagnosis method based on dynamic operating condition matching and hierarchical contribution analysis.

[0032] Figure 2 The flowchart illustrates a method for diagnosing energy consumption in coking processes based on dynamic operating condition matching and hierarchical contribution analysis, provided by this invention.

[0033] Figure 3 This is a schematic diagram of the data structure of the coking process within the rolling diagnosis cycle of a coking process energy consumption diagnosis method based on dynamic operating condition matching and hierarchical contribution analysis provided by the present invention.

[0034] Figure 4 The present invention provides a diagnostic flowchart for a coking process energy consumption diagnosis method based on dynamic operating condition matching and hierarchical contribution analysis.

[0035] Figure 5 The diagram shows the influencing factors of a coking process energy consumption diagnosis method based on dynamic operating condition matching and hierarchical contribution analysis provided by this invention. Detailed Implementation

[0036] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0037] This invention provides a method for diagnosing energy consumption in coking processes based on dynamic operating condition matching and hierarchical contribution analysis, such as... Figure 1-2 As shown, it includes the following steps: S1. Based on the current diagnostic time t The end time is [time value], and the construction time is [duration value]. Δt The rolling diagnostic cycle collects historical production and energy data for the coking process within that cycle.

[0038] The formula for calculating the rolling diagnostic cycle is: [ t-Δt,t ] The rolling diagnostic cycle follows a preset update step. δt Rolling updates, among which δt Less than or equal to Δ t Δ t It can be configured to be hour, shift, production batch or natural day according to on-site management and data conditions.

[0039] like Figure 3As shown, the historical production and energy data of the coking process during the rolling diagnostic cycle include at least the following: coal moisture content, coke oven gas calorific value, blast furnace gas calorific value, carbonization time, vertical running temperature, coke production, coking coal consumption, non-coking coal consumption, blast furnace gas consumption, coke oven gas consumption, low-pressure steam consumption, electricity consumption, nitrogen consumption, industrial and domestic water consumption, high-pressure steam and low-pressure steam recovery, external supply of coke oven gas, tar production, and crude benzene production.

[0040] S2. Based on the collected historical production and energy data, a coking process production operation database is constructed. Other parameters already configured on-site can be integrated into the production operation data database as extended diagnostic factors. For input, output, and process parameters with significant process lags, time alignment is performed based on carbonization time or preset process delays; cumulative quantities are incremented within the diagnostic cycle, and continuous variables are calculated based on data attributes to determine the cycle mean, integral value, or characteristic value, thereby forming a record of actual operating conditions corresponding to the current diagnostic cycle.

[0041] The coking process production operation database is used not only for calculating energy consumption in the current performance cycle, but also for forming a comparable sample library composed of multiple historical operating condition segments within a preset historical time range. The preset historical time range is 3 to 36 months before the end of the current diagnostic cycle. Among them, 3 months is used as the lower limit of the historical sample window to ensure a sufficient number of normal production conditions and reduce the impact of short-term fluctuations on the benchmark selection. The historical production data of the previous 12 months is preferred to cover the periodic changes in raw materials, environment, and production load. When the effective sample is insufficient, it can be gradually extended backward, but the maximum period shall not exceed 36 months.

[0042] The historical production and energy data of the coking process collected during the rolling diagnostic cycle are processed by unifying time windows, removing duplicate records, marking missing values, identifying outliers, and unifying units. For cumulative data, the increment within the rolling diagnostic cycle is calculated. For flow, temperature, and calorific value data, the cumulative value, mean, or characteristic value of the rolling diagnostic cycle is calculated based on the data attributes, and then converted into unit coke index with the coke production of the rolling diagnostic cycle as the denominator. Operating condition records with zero coke production, missing metering, exceeding reasonable range, incomplete time windows, or violating preset material and energy constraints are marked as invalid.

[0043] S3. Based on the coking process production operation database, construct a comprehensive energy consumption model per unit of coke, with the following calculation formula:

[0044] In the formula: E coking The unit coke comprehensive energy consumption of the coking process during the rolling diagnostic cycle is expressed in kg / t. Ecoal , E' cog E Tar , E ber , E coke , E cog , E bfg , E dljz , E steam , E ele The corresponding standard coal equivalents (kg) for washed clean coal, coke oven gas, tar, benzene, coke, coke oven gas and blast furnace gas, power medium, recovered steam, and self-generated electricity during the diagnostic period are: p coke The figure represents the coke production in tons during the diagnostic period. Coke oven gas is a byproduct of the coking process that can be supplied externally or recycled; it is typically deducted from overall energy consumption. Coke oven gas is used as fuel and is considered an energy consumption item.

[0045] According to the energy conversion standard and by-product energy recovery conversion rules, the standard coal equivalent of input energy, recovered energy, and by-product energy during the diagnostic period are calculated separately, and the comprehensive energy consumption per unit of coke is obtained by combining the period's coke production. The input energy includes coking coal, non-coking coal, blast furnace gas, coke oven gas, steam, electricity, nitrogen, and water, etc.; the recovered items include high-pressure steam, low-pressure steam, externally supplied coke oven gas, tar, and crude benzene, etc.

[0046] S4. For example Figure 4 As shown, in accordance with the principles of operating conditions within the same production boundary and satisfying preset parameter range, valid data, and priority of low energy consumption, historical operating condition samples with the same time window length as the current actual performance cycle and satisfying preset constraints are extracted from the coking process production operation database. Samples that meet preset constraints, such as coke oven production line, equipment status, production load, coal moisture, coke oven gas calorific value, blast furnace gas calorific value, carbonization time, and vertical temperature, are screened out, and operating condition segments that are shut down, under maintenance, have missing main data, or are obviously abnormal are removed.

[0047] Based on each historical working condition sample, a historical working condition feature vector containing preset index parameters is constructed. Based on the current actual working condition, a current performance cycle working condition feature vector containing the same preset index parameters is constructed. The historical working condition feature vector and the current performance cycle working condition feature vector are compared. Historical working conditions that meet the preset similarity conditions are selected from the historical working condition samples as comparable historical working conditions.

[0048] S5. Calculate the unit coke comprehensive energy consumption for each comparable historical operating condition using the unit coke comprehensive energy consumption model, and sort them from low to high. Select the comparable historical operating condition with the lowest energy consumption value or within a preset percentile interval and stable operation as the advanced benchmark operating condition. The advanced benchmark operating condition is not a fixed date in advance, but a low-energy-consumption historical operating condition segment dynamically determined for each diagnostic cycle after data governance and operating condition similarity screening.

[0049] After obtaining the advanced benchmark operating conditions, the unit coke comprehensive energy consumption for the current performance period is calculated using the unit coke comprehensive energy consumption model. E actual Comprehensive energy consumption per unit of coke corresponding to advanced benchmark operating conditions E base According to Δ E = E actual - E base The difference in unit coke comprehensive energy consumption between the current performance period and the advanced benchmark operating condition is calculated; where Δ E If the value is greater than zero, it is determined that there is energy-saving potential in the current operating condition, and step S6 is continued; Δ E If the value is less than or equal to zero, it is determined that there is no energy-saving potential in the current operating condition, and the next step is stopped.

[0050] The current performance cycle and the advanced benchmark operating conditions use the same diagnostic cycle length, production boundary, standard coal conversion factor, and unit coke conversion caliber, making the input energy, recovered energy, and by-product data of the two comparable.

[0051] The preset parameters include at least production load, coal moisture content, coke oven gas calorific value, blast furnace gas calorific value, carbonization time, and vertical travel temperature.

[0052] The preset similarity conditions can be determined by parameter difference, weighted distance, clustering results or rule matching, and can be configured according to the on-site coke oven production organization and data availability.

[0053] Comparable historical operating conditions refer to historical operating condition segments within a preset historical time range that meet similarity conditions to the current performance cycle in terms of production boundaries, equipment status, diagnostic cycle length, production load, and key process parameters, and whose data is complete and valid, without triggering shutdowns, maintenance, abnormal fluctuations, or data missing markers. Among comparable historical operating conditions, the record with the lowest unit coke comprehensive energy consumption and stable production status is selected as the advanced benchmark operating condition, sorted from low to high. When multiple similar low-energy-consumption candidate operating conditions exist, the operating condition segment with smaller data fluctuations and better stability within the low-energy-consumption quantile interval can be selected as the benchmark. The input energy, recovered energy, by-products, process parameters, and unit energy consumption of this advanced benchmark operating condition form a benchmark parameter set for subsequent benchmarking and diagnostics. The advanced benchmark operating condition is updated continuously with the diagnostic cycle, and its determination process includes four stages: historical operating condition segment acquisition, abnormal data removal, production condition similarity screening, and low-energy-consumption sample selection.

[0054] S6. For example Figure 5 As shown, according to energy income and expenditure categories, the unit coke comprehensive energy consumption under the current actual operating condition and the advanced benchmark operating condition determined from comparable historical operating conditions is decomposed into four primary influencing factors: fuel consumption, energy conversion difference, electricity consumption, and waste energy recovery. Fuel consumption mainly reflects the impact of changes in the consumption and calorific value of fuel media such as blast furnace gas and coke oven gas on energy consumption; energy conversion difference mainly reflects the impact of changes in the income and expenditure of standard coal equivalent, such as coking coal, non-coking coal, coke, tar, crude benzene, and coke oven gas generation and external supply, on energy consumption; electricity consumption mainly reflects the impact of changes in electricity consumption for transportation, dust removal, fans, pumps, and recovery systems on energy consumption; waste energy recovery mainly reflects the impact of changes in the recovery of high-pressure steam, low-pressure steam, dry quenching coke, and waste heat from riser pipes on energy consumption.

[0055] Based on the unit coke comprehensive energy consumption, the energy consumption components of the primary influencing factors are calculated for both the current performance cycle operating conditions and the advanced benchmark operating conditions. Energy consumption items are counted as positive values, while energy recovery items and by-product deduction items are counted as negative values ​​in the unit coke comprehensive energy consumption model. The difference between the energy consumption components of the same primary influencing factor under the current performance cycle operating conditions and the advanced benchmark operating conditions is determined as the primary influence value of that primary influencing factor. The algebraic sum of all primary influence values ​​is determined as the total change in energy consumption.

[0056] The total change in unit coke comprehensive energy consumption between the current actual operating condition and the advanced benchmark operating condition is calculated based on the energy balance relationship. The ratio of each primary influence value to the total change in unit coke comprehensive energy consumption is calculated to obtain the contribution of each primary influence factor.

[0057] The factors are sorted from largest to smallest based on the absolute value of the primary impact value, and the diagnosis of whether each primary impact factor leads to an increase or decrease in energy consumption in the coking process is determined based on the sign of the primary impact value.

[0058] No. i The first-level influence value of each first-level influencing factor I i The calculation formula is as follows:

[0059] Total change in comprehensive energy consumption per unit of coke ΔC The calculation formula is as follows:

[0060] The formula for calculating the contribution of primary influencing factors is as follows:

[0061] In the formula: or i For the first i The contribution of each primary influencing factor to the change in energy consumption in the coking process; C i,actual For the first i The energy consumption components of each primary influencing factor are uniformly converted according to the positive and negative directions in the unit coke comprehensive energy consumption model under the current actual performance cycle conditions. C i,base For the first i The energy consumption components of each primary influencing factor are calculated using the same caliber under advanced benchmark conditions determined from comparable historical operating conditions. ΔC This represents the total change in energy consumption during the coking process; m The number of primary influencing factors; k The subscript is used for summation.

[0062] The four primary influencing factors—fuel consumption, energy conversion deficit, electricity consumption, and waste energy recovery—are formed by merging the energy input, output, and recovery deduction items in the unit coke comprehensive energy consumption model according to their energy income and expenditure attributes. Each energy item in the unit coke comprehensive energy consumption model is included in the corresponding primary influencing factor according to its energy income and expenditure category.

[0063] The decomposition of primary influencing factors in this invention refers to expressing the change in the comprehensive energy consumption per unit of coke between the current actual operating conditions and the advanced benchmark operating conditions determined from comparable historical operating conditions as the algebraic sum of the primary influence values ​​corresponding to each primary influencing factor, based on the unit coke comprehensive energy consumption model and the energy balance relationship of the coking process. The primary influence value is the difference in energy consumption components of the same primary influencing factor under the current actual operating conditions and the advanced benchmark operating conditions; no further conversion of this difference is required.

[0064] Figure 5The secondary parameters shown are used to illustrate the parameter hierarchy in the energy consumption diagnosis of the coking process and its correlation with the primary influencing factors. The specific secondary parameters involved in the diagnosis and their corresponding primary influencing factors are determined based on the energy expenditure relationship in the unit coke comprehensive energy consumption model, on-site data definitions, and sensitivity analysis screening results.

[0065] S7. Select the primary influencing factors whose absolute contribution value is greater than the preset threshold, obtain the values ​​of the secondary parameters of each primary influencing factor under the current performance cycle conditions and the advanced benchmark conditions determined from comparable historical conditions; calculate the diagnostic parameters of each secondary parameter based on these values, and output the diagnostic conclusions and adjustment suggestions for each secondary parameter.

[0066] Secondary parameters for fuel consumption include at least coal moisture content, coke oven gas calorific value, blast furnace gas calorific value, carbonization time, and vertical operating temperature. Secondary parameters for poor energy conversion include coal moisture content, coal volatile matter, coal ash content, carbon monoxide content in circulating gas, coking time, and combustion chamber temperature, which can correspond to changes in coal consumption, coke oven gas generation, and by-product generation, respectively. Secondary parameters for electricity consumption include load proxy parameters such as cycle coal consumption, tar production, and washed coal moisture content, which can be further refined into operable parameters such as the operating time and load rate of dust removal fans, conveyor belts, pumps, and recovery system motors when equipment-level data is available. Secondary parameters for waste energy recovery include vertical operating temperature, coal moisture content, cold coke temperature, coke volatile matter, riser outlet temperature, high-pressure steam enthalpy, low-pressure steam enthalpy, and carbonization time.

[0067] The aforementioned secondary parameters include both field-adjustable and verifiable parameters such as coal quality, calorific value, temperature, time, and steam enthalpy, as well as proxy parameters used to characterize changes in equipment operating load and production organization under current data conditions. When equipment-level operating data is added to the field data interface, the secondary parameters can be further expanded.

[0068] For each secondary parameter, obtain the value of the secondary parameter under the current performance cycle working condition and the value under the advanced benchmark working condition determined from comparable historical working conditions, calculate the secondary relative change rate and secondary impact value, and sort them from largest to smallest according to the absolute value of the secondary contribution, output the top-ranked parameters and their change direction; by comparing the relationship between the increase or decrease of the secondary parameter and the change direction of the corresponding primary influencing factor, form a parameter adjustment prompt for the field.

[0069] The diagnostic conclusions for each secondary parameter include the value of the secondary parameter in the current performance period, the value of the secondary parameter under the advanced benchmark condition determined in comparable historical conditions, the relative change rate of the secondary parameter, the secondary contribution, the ranking results based on the absolute value of the secondary contribution, and the diagnostic conclusion that the secondary parameter causes the corresponding primary influencing factor to increase or decrease.

[0070] The formula for calculating the second-order relative rate of change is as follows:

[0071] In the formula: r j For the first j The relative rate of change of each second-order parameter; x j,actual For the first j The values ​​of each secondary parameter in the current performance period; x j,base For the first j The values ​​of each secondary parameter are determined under the advanced benchmark condition in comparable historical operating conditions.

[0072] when When the value is 0, the second-order relative change rate is not calculated, and a prompt indicating that the baseline value is zero is output. The second-order relative change rate is only used to characterize the deviation of the actual value of the second-order parameter from the advanced baseline value, and is not involved in the calculation of the second-order parameter sensitivity, energy consumption impact change, second-order contribution, ranking results, and the direction of influence of the second-order parameter.

[0073] The formula for calculating the sensitivity of each secondary parameter is as follows:

[0074]

[0075] In the formula: S j For the second-level parameters, the first j Sensitivity corresponding to each variable value; E j For the second-level parameters, the first j Energy consumption values ​​corresponding to each variable value; P j For the second-level parameters, the first j One variable value; E o This represents the energy consumption value corresponding to the basic variable value under the secondary parameters; P o The reference variable value under the second-level parameter; S average This represents the average sensitivity of the variable values ​​under the second-order parameter, i.e., the sensitivity coefficient. n This represents the total number of variables under the second-level parameters; Based on the secondary parameters and the comparable historical operating data corresponding to each secondary parameter, a multiple linear regression curve is used to establish the relationship between the secondary parameters and the corresponding primary influencing factors. The calculation formula is as follows:

[0076] In the formula:y , where is the target variable, representing the energy consumption impact value of the primary influencing factor corresponding to the secondary parameter; β 0 represents the intercept term; β 1. β 2、…、 β n These are the regression coefficients obtained by fitting a sample of comparable historical operating conditions. x 1. x 2、…、 x j These are the secondary parameters selected for modeling after sensitivity analysis. Sensitivity analysis involves calculating the sensitivity of each secondary parameter at different value points to obtain the average sensitivity. Then, the parameters are sorted from largest to smallest by the absolute value of the average sensitivity, and secondary parameters that exceed the preset threshold or rank highly are selected for subsequent fitting modeling and contribution calculation. j The number of secondary parameters involved in the modeling; e This is the error term; The formula for calculating the secondary contribution is as follows:

[0077] In the formula, or j For the first j The contribution of each secondary parameter; Δ y j For the first j The change in energy consumption caused by the transformation of a secondary parameter from its value under the advanced benchmark condition determined in comparable historical operating conditions to its value under the current actual operating condition; Δ y j The sign indicates the direction of the effect of the secondary parameter on energy consumption, whether it increases or decreases.

[0078] In this invention, the primary and secondary relative change rates are calculated using the same methodology, both employing the corresponding index values ​​from the advanced benchmark operating conditions determined from comparable historical operating conditions as the comparison base. Specifically, the primary relative change rate characterizes the deviation of the primary factor's energy consumption component from the advanced benchmark operating condition, while the secondary relative change rate characterizes the deviation of the secondary parameter's actual value from the advanced benchmark value. Both the primary and secondary relative change rates are auxiliary characterization indicators in the diagnostic results and are not involved in the calculation of primary contribution, secondary contribution, ranking results, or the direction of energy consumption influence.

[0079] It should be noted that this invention uses a unified calculation method only for the first-level relative change rate and the second-level relative change rate, both calculated by subtracting the benchmark value from the actual value and then dividing by the benchmark value. The first-level and second-level contribution rates are calculated separately for different levels of diagnostic objects: the first-level contribution rate is determined by the ratio of the first-level impact value to the total change in comprehensive energy consumption per unit of coke, used to characterize the impact of each first-level influencing factor on the total change in energy consumption; the second-level contribution rate is determined by the ratio of the absolute value of the energy consumption impact change caused by the second-level parameter to the sum of the absolute values ​​of the energy consumption impact changes of each second-level parameter, used to characterize the relative magnitude of the influence of each second-level parameter. The direction of the energy consumption increase or decrease caused by the second-level parameter is determined according to the energy consumption impact change Δ... y j The positive or negative sign is determined.

[0080] If the current secondary parameters are in good condition, the output suggests that no adjustment is needed; when there are secondary parameters that need to be optimized, the output suggests the variables to be adjusted, the suggested adjustment amount, and the expected changes in the corresponding primary influencing factors and the comprehensive energy consumption per unit of coke.

[0081] Generate a structured diagnostic report based on a preset update step size. The report includes the start and end times of the diagnostic cycle, the comprehensive energy consumption per unit of coke in the current performance cycle, the comprehensive energy consumption per unit of coke under advanced benchmark conditions, energy saving potential, operating condition similarity, ranking of primary factors such as fuel consumption / energy conversion difference / electricity consumption / residual energy recovery, and the top-ranked secondary parameters.

[0082] Based on the top-ranked secondary parameters, suggestions corresponding to the parameter categories are output. For example, when coal moisture, volatile matter, or ash content deviates from the advanced benchmark conditions and has a significant impact on energy conversion or fuel consumption, suggestions for stabilizing the quality of coal fed into the furnace and coal blending conditions are output; when coke oven gas calorific value, blast furnace gas calorific value, carbonization time, coking time, or vertical travel temperature deviates from the advanced benchmark conditions, suggestions for reviewing the heating system are output; when parameters related to high-pressure steam, low-pressure steam, riser outlet temperature, or steam enthalpy deviate from the advanced benchmark conditions, suggestions for checking the operation status and metering integrity of waste heat recovery are output; when electricity consumption is high and related to differences in coal consumption, tar production, washed coal moisture, or equipment operating load, suggestions for reviewing the operation status of conveying, dust removal, fans, pumps, and recovery systems are output.

[0083] The adjusted secondary parameter values ​​are input into the prediction model to calculate the adjusted primary influencing factors and the comprehensive energy consumption per unit of coke. After the recommendations are implemented, subsequent diagnostic cycle data are collected according to the preset update step size and compared with the difference between the actual operating condition and the baseline operating condition before implementation. When the energy-saving potential has not decreased or the main influencing factors have not improved, the comparable operating condition screening, advanced baseline operating condition identification, and contribution diagnosis are re-executed, forming a closed loop process of cycle data governance, dynamic operating condition matching, hierarchical diagnosis, and recommendation output.

[0084] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for diagnosing energy consumption in coking processes based on dynamic operating condition matching and hierarchical contribution analysis, characterized in that, Includes the following steps: S1. Based on the current diagnostic time t The end time is [time value], and the construction time is [duration value]. Δt The rolling diagnostic cycle collects historical production and energy data for the coking process within that cycle. S2. Based on the collected historical production and energy data, construct a production operation database for the coking process; S3. Based on the coking process production operation database, construct a comprehensive energy consumption model per unit of coke; S4. Extract historical operating condition samples from the coking process production operation database that have the same time window length as the current performance cycle and meet the preset constraints. Construct a historical operating condition feature vector containing preset index parameters based on each historical operating condition sample. Construct a current performance cycle operating condition feature vector containing the same preset index parameters based on the current performance operating condition. Compare the historical operating condition feature vector with the current performance cycle operating condition feature vector. Select historical operating conditions that meet the preset similarity conditions from the historical operating condition samples as comparable historical operating conditions. S5. Calculate the unit coke comprehensive energy consumption of each comparable historical working condition through the unit coke comprehensive energy consumption model, and sort them from low to high according to the unit coke comprehensive energy consumption. Select the comparable historical working condition with the lowest energy consumption value or within the preset quantile range and with stable operation as the advanced benchmark working condition. S6. According to the energy income and expenditure category, the unit coke comprehensive energy consumption under the current performance cycle operating condition and the advanced benchmark operating condition determined from the comparable historical operating conditions is decomposed into several first-level influencing factors. Based on the unit coke comprehensive energy consumption, the energy consumption component of the first-level influencing factors under the current performance cycle operating condition and the advanced benchmark operating condition is calculated respectively. The difference between the energy consumption components of the first-level factors under the current performance cycle operating condition and the advanced benchmark operating condition is determined as the first-level influence value of the first-level influencing factor. The total change in unit coke comprehensive energy consumption between the current actual operating condition and the advanced benchmark operating condition is calculated based on the energy balance relationship. The ratio of each primary influence value to the total change in unit coke comprehensive energy consumption is calculated to obtain the contribution of each primary influence factor. The factors are sorted from largest to smallest according to the absolute value of the first-level impact value, and the diagnosis of whether each first-level impact factor leads to an increase or decrease in energy consumption in the coking process is determined based on the sign of the first-level impact value. S7. Select the primary influencing factors whose absolute contribution value is greater than the preset threshold, obtain the values ​​of the secondary parameters of each primary influencing factor under the current performance cycle conditions and the advanced benchmark conditions determined from comparable historical conditions; calculate the diagnostic parameters of each secondary parameter based on these values, and output the diagnostic conclusions and adjustment suggestions for each secondary parameter.

2. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, The formula for calculating the rolling diagnostic cycle in step S1 is: [ t-Δt,t ] The rolling diagnostic cycle follows a preset update step. δt Rolling updates, among which δt Less than or equal to Δt .

3. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, The historical production and energy data of the coking process collected during the rolling diagnostic cycle are processed by unifying time windows, removing duplicate records, marking missing values, identifying outliers, and unifying units. For cumulative data, the increment within the rolling diagnostic cycle is calculated. For flow rate, temperature, and calorific value data, the cumulative value, mean, or characteristic value of the rolling diagnostic cycle is calculated based on the data attributes, and then converted into unit coke index with the coke production of the rolling diagnostic cycle as the denominator. Records of working conditions with zero coke production, missing metering, exceeding reasonable range, incomplete time windows, or violating preset material and energy constraints are marked as invalid.

4. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, In step S3, a comprehensive energy consumption model per unit of coke is constructed based on the coking process production operation database, and the calculation formula is as follows: In the formula: E coking The unit coke comprehensive energy consumption of the coking process during the rolling diagnostic cycle is expressed in kg / t. E coal , E' cog E Tar , E ber , E coke , E cog , E bfg , E dljz , E steam , E ele The corresponding standard coal equivalents (kg) for washed clean coal, coke oven gas, tar, benzene, coke, coke oven gas and blast furnace gas, power medium, recovered steam, and self-generated electricity during the diagnostic period are: p coke The coke production during the diagnostic cycle is expressed in tons (t).

5. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, After obtaining the advanced benchmark operating conditions in step S5, the unit coke comprehensive energy consumption for the current performance period is calculated using the unit coke comprehensive energy consumption model. E actual Comprehensive energy consumption per unit of coke corresponding to advanced benchmark operating conditions E base According to Δ E = E actual - E base The difference in unit coke comprehensive energy consumption between the current performance period and the advanced benchmark operating condition is calculated; where Δ E If the value is greater than zero, it is determined that there is energy-saving potential in the current operating condition, and step S6 is continued; Δ E If the value is less than or equal to zero, it is determined that there is no energy-saving potential in the current operating condition, and the next step is stopped.

6. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, In step S6, the first i The first-level influence value of each first-level influencing factor I i The calculation formula is as follows: Total change in comprehensive energy consumption per unit of coke ΔC The calculation formula is as follows: The formula for calculating the contribution of primary influencing factors is as follows: In the formula: η i For the first i The contribution of each primary influencing factor to the change in energy consumption in the coking process; C i,actual For the first i The energy consumption components of each primary influencing factor are uniformly converted according to the positive and negative directions in the unit coke comprehensive energy consumption model under the current actual performance cycle conditions. C i,base For the first i The energy consumption components of each primary influencing factor are calculated using the same caliber under advanced benchmark conditions determined from comparable historical operating conditions. ΔC This represents the total change in energy consumption during the coking process; m The number of primary influencing factors; k Subscripts are used for summation.

7. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 6, characterized in that, The diagnostic conclusions of each primary influencing factor output in step S6 include the primary factor energy consumption component under the current actual performance cycle conditions, the primary factor energy consumption component under the advanced benchmark conditions determined from comparable historical conditions, the primary relative change rate, the primary contribution, the ranking results, and the diagnostic conclusions that the primary influencing factors cause an increase or decrease in the energy consumption of the coking process. The formula for calculating the relative rate of change is as follows: In the formula: r i For the first i The relative rate of change of each primary influencing factor is used to characterize the degree of deviation of the energy consumption component of the primary factor from the advanced benchmark under the current actual operating conditions. when When this happens, the first-order relative rate of change is not calculated, and a prompt indicating that the baseline value is zero is output. when Ii When the value is greater than 0, it is determined that this primary influencing factor leads to an increase in energy consumption in the coking process; when... Ii When <0, it is determined that this primary influencing factor leads to a reduction in energy consumption in the coking process; when Ii When the value is 0, it is determined that this primary influencing factor has no significant impact on the energy consumption changes in the coking process.

8. The energy consumption diagnosis method for coking processes based on dynamic operating condition matching and hierarchical contribution analysis according to claim 1, characterized in that, The diagnostic conclusions for each secondary parameter include the value of the secondary parameter in the current performance period, the value of the secondary parameter under the advanced benchmark condition determined in comparable historical conditions, the relative change rate of the secondary parameter, the contribution of the secondary parameter, the ranking results based on the absolute value of the contribution of the secondary parameter, and the diagnostic conclusion that the secondary parameter causes the corresponding primary influencing factor to increase or decrease. The formula for calculating the second-order relative rate of change is as follows: In the formula: r j For the first j The relative rate of change of each of the second-order parameters; x j,actual For the first j The values ​​of each secondary parameter in the current performance period; x j,base For the first j The values ​​of each secondary parameter under the advanced benchmark condition determined in comparable historical operating conditions; when x j,base When =0, the second-order relative rate of change is not calculated, and a prompt indicating that the baseline value is zero is output. The formula for calculating the sensitivity of each secondary parameter is as follows: In the formula: S j For the second-level parameters, the first j Sensitivity corresponding to each variable value; E j For the second-level parameters, the first j Energy consumption values ​​corresponding to each variable value; P j For the second-level parameters, the first j One variable value; E o This represents the energy consumption value corresponding to the basic variable value under the secondary parameters; P o The reference variable value under the second-level parameter; S average This represents the average sensitivity of the variable values ​​under the second-order parameters, i.e., the sensitivity coefficient. n This represents the total number of variables under the second-level parameters; Based on the secondary parameters and the comparable historical operating data corresponding to each secondary parameter, a multiple linear regression curve is used to establish the relationship between the secondary parameters and the corresponding primary influencing factors. The calculation formula is as follows: In the formula: y , where is the target variable, representing the energy consumption impact value of the primary influencing factor corresponding to the secondary parameter; β 0 represents the intercept term; β 1. β 2、…、 β n These are the regression coefficients obtained by fitting a sample of comparable historical operating conditions. x 1. x 2、…、 x j These are the secondary parameters that were selected for modeling after sensitivity analysis; j The number of secondary parameters involved in the modeling; ε This is the error term; The formula for calculating the secondary contribution is as follows: In the formula, η j For the first j The contribution of each secondary parameter; Δ y j For the first j The change in energy consumption caused by the change of the value of each secondary parameter from the advanced benchmark condition determined in comparable historical conditions to the value under the current actual performance period condition. Δ y j The sign indicates the direction of the effect of the secondary parameter on energy consumption, whether it increases or decreases.

Images