Method for evaluating quality of wood chips for kraft pulping

Abstract

The application discloses a method for evaluating the quality of wood chips of a kraft pulp raw material, and the steps are as follows: obtaining qualified wood chips by screening a plurality of detection samples, and determining the wood chip qualification rate of the detection samples; collecting data by visual detection and calculating adsorption uniformity and uniformity scores; determining pulp yield and fiber length by kraft pulping of the qualified wood chips, and synchronously calculating qualification rate scores, yield scores and length scores; obtaining a total wood chip quality score of the detection samples by weighted summation after optimizing dynamic weight coefficients based on a deep reinforcement learning algorithm; and performing cumulative probability analysis based on the total wood chip quality scores of all the detection samples to grade the quality of the raw material wood chips. The introduction of multi-dimensional scoring indexes and the optimization of dynamic weight coefficients improve the accuracy of wood chip quality evaluation, solve the economic benefit problems caused by the quality of raw materials in the industrial production process of wood chips, and provide data basis for raw material procurement.

Description

Technical Field

[0001] This invention relates to the field of papermaking technology, and in particular to a method for evaluating the quality of wood chips used as raw material in sulfate pulping. Background Technology

[0002] In sulfate pulping, raw material costs account for approximately 60%-70% of total production costs, making it a key factor affecting the economic benefits of enterprises. Currently, sulfate pulp plants generally purchase log chips for the production process. Due to the requirements of the production process on raw material specifications, most pulp plants in China mainly use the good wood chip rate as the core evaluation indicator for raw material procurement. This is mainly because a higher good wood chip rate means a higher proportion of wood chips that can be utilized during production, less of which needs to be screened out, and thus lower raw material costs. However, this indicator only reflects the degree of qualification of wood chip size specifications, while ignoring the pulping performance of the wood chips after cooking. Therefore, in addition to the specifications of wood chips, the pulping performance of wood chips is also an important consideration. For pulp products, the properties of wood chips are a key factor affecting the quality of pulp products. When the same tree species is used to prepare sulfate pulp under different regions, different ages and different growth environments, its pulping performance will also vary significantly. In the procurement process, for wood chips of the same tree species, there is a lack of wood chip quality assessment criteria, and they are often treated with the same principles. However, purchasing inferior wood chips will not only cause production cost losses, but may also cause production instability and affect product quality. Summary of the Invention

[0003] In view of this, the present invention proposes an evaluation method for the quality of wood chips used as raw materials for sulfate pulping, which adopts a multi-dimensional scoring index for weighted scoring to improve the accuracy of wood chip quality assessment.

[0004] The technical solution of this invention is implemented as follows: A method for evaluating the quality of wood chips used in sulfate pulping includes the following steps: Step S1: Collect a number of raw wood chips and divide them into multiple groups of test samples. After screening the test samples, obtain qualified wood chips and determine the qualified rate of the wood chips in the test samples. Step S2: Collect multidimensional thickness data of qualified wood chips through visual inspection, calculate adsorption uniformity, and calculate the uniformity score of the test sample based on adsorption uniformity. Step S3: Pulp qualified wood chips using the sulfate method, determine the pulp yield and fiber length, and calculate the pass rate score, yield score and length score of the test sample using the standardized score scoring method. Step S4: Based on the deep reinforcement learning algorithm, optimize the dynamic weight coefficients of uniformity score, pass rate score, yield score and length score; Step S5: Based on the dynamic weighting coefficient, the uniformity score, pass rate score, yield score and length score are weighted and summed to obtain the total quality score of the wood chips of the test sample. Step S6: Perform cumulative probability analysis based on the total quality score of all tested wood chips to classify the quality of raw wood chips into grades.

[0005] Preferably, step S1 includes the following steps: Step S11: Purchase raw wood chips of the same tree species and divide them into multiple groups of test samples. Screen the test samples according to the TAPPI standard. Step S12: Screening to obtain raw wood chips with a diameter range of 9.5mm to 40mm as qualified wood chips; Step S13: Calculate the wood chip pass rate of the test sample based on the quality of qualified wood chips and raw wood chips.

[0006] Preferably, step S2 includes the following specific steps: Step S21: Evenly distribute measurement points on the upper surface, lower surface, edge, and center of the qualified wood chips; Step S22: Use a convolutional neural network vision inspection system to obtain the original data of the wood chip thickness at the measurement points; Step S23: Based on the original wood chip thickness data, calculate the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution; Step S24: Calculate the adsorption uniformity based on the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution, and obtain the uniformity score of the test sample using the inverse index scoring.

[0007] Preferably, the formula for calculating the adsorption uniformity is:

[0008] The formula for calculating the uniformity score is:

[0009] in For uniform adsorption, For thickness standard deviation, The average thickness Due to extremely poor thickness, For thickness distribution skewness, To fix the correction factor, The exponential decay coefficient is... Let be the uniformity score of the i-th test sample. For the adsorption uniformity of the i-th test sample, The minimum adsorption uniformity among all tested samples, This represents the maximum adsorption uniformity across all tested samples.

[0010] Preferably, step S3, which involves pulping qualified wood chips using the sulfate process and determining the pulp yield and fiber length, includes the following specific steps: Step S31: After drying the qualified wood chips, put them into the digester and add cooking liquor to the digester. Set the corresponding cooking conditions to cook and pulp. Step S32: Wash the pulp obtained after cooking and calculate the pulp yield based on the sample's oven-dry weight. Step S33: Take the cooked and washed pulp and use a fiber analyzer to measure the fiber length.

[0011] Preferably, step S3, which uses a standardized score scoring method to calculate the pass rate score, yield score, and length score, includes the following specific steps: Calculate the mean and standard deviation of the wood chip pass rate A, pulp yield Y, and fiber length F for the tested samples respectively; The standardized scoring method is used for scoring, and the scoring formula is as follows:

[0012]

[0013]

[0014] in These are the pass rate score, yield score, and length score for the i-th tested sample, respectively. These represent the wood chip pass rate, pulp yield, and fiber length of the i-th test sample, respectively. To detect the mean wood chip pass rate, mean pulp yield, and mean fiber length of the corresponding samples, The standard deviations of the wood chip pass rate, pulp yield, and fiber length corresponding to the test samples were determined.

[0015] Preferably, step S4 includes the following specific steps: Step S41: Construct the state space S of the pulping production environment. The state space is composed of the average wood chip qualification rate, the average pulp yield, the average fiber length, the average adsorption uniformity, the cooking energy consumption, the slag screening rate, and the tensile strength of the pulp. Step S42: Construct an action space M, which outputs a 4-dimensional dynamic weight coefficient. Design a multi-objective reward function R to comprehensively reward the improvement of slurry yield, strength attainment, energy consumption reduction, and adsorption uniformity. Step S43: Using the DDPG dual network structure, the Actor network outputs actions based on the state space, and the Critic network evaluates the merits of the actions based on the reward function and performs feedback iteration. Step S44: Process the original weights output by the Actor network through softmax normalization constraints, and output the optimal dynamic weight coefficients.

[0016] Preferably, the expression for the multi-objective reward function is:

[0017] in The reward value for the current state action pair. This represents the actual slurry yield. To achieve the target slurry yield, For the actual tensile strength of the slurry, To achieve the target tensile strength of the slurry, For energy consumption, For maximum energy consumption, For the slag screening rate, To achieve the maximum slag screening rate, As a benchmark for adsorption uniformity, For the adsorption uniformity of the i-th test sample, These are the weighting coefficients.

[0018] Preferably, the formula for calculating the total quality score of the wood chips is:

[0019] in The total score for the quality of wood chips in the i-th test sample. These represent the pass rate score, yield score, length score, and uniformity score for the i-th tested sample, respectively. These are dynamic weighting coefficients.

[0020] Preferably, the grading rule in step S6 is as follows: Wood chips with a cumulative probability of ≤20% in the total quality score are classified as Grade 5 wood chips. Wood chips with a cumulative probability of 20% or less of the total quality score being ≤40% are classified as Grade 4 wood chips. Wood chips with a cumulative probability of 40% or less of the total quality score and ≤60% are classified as Grade 3 wood chips. Wood chips with a cumulative probability of 60% or less of the total quality score and ≤80% are classified as Grade II wood chips. A wood chip with a cumulative overall quality score of less than 80% is considered a first-class wood chip.

[0021] Compared with the prior art, the beneficial effects of the present invention are: This invention discloses a method for evaluating the quality of wood chips used in sulfate pulping. After obtaining wood chips from the same tree species, these chips are divided into several test samples. Each group of test samples is sieved to obtain qualified wood chips, and the qualified rate is determined. After sieving, multi-dimensional thickness data of the qualified wood chips is collected using a visual inspection device, and adsorption uniformity is calculated based on this data. Simultaneously, the sieved qualified wood chips can be pulped using the sulfate process, and the pulp yield and fiber length are measured. The qualified rate, adsorption uniformity, pulp yield, and fiber length are used as scoring indicators, and a score is calculated for each indicator. A deep reinforcement learning algorithm is introduced to adaptively match optimal weights based on the real-time state of pulping production, significantly improving the accuracy, stability, and production adaptability of wood chip quality scoring. Finally, based on the scores and weights, the total wood chip quality score for each test sample can be calculated. After performing cumulative probability analysis on the total wood chip quality scores of all test samples, the quality grade of each test sample can be determined. This addresses the economic benefits caused by raw material quality in the industrial production of wood chips and provides data support for raw material procurement. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a method for evaluating the quality of wood chips used in sulfate pulping according to the present invention. Figure 2 This is a flowchart of step S1 of a method for evaluating the quality of wood chips used in sulfate pulping according to the present invention. Figure 3 This is a flowchart of step S2 of a method for evaluating the quality of wood chips used in sulfate pulping according to the present invention. Figure 4 This is a flowchart of step S3 in a method for evaluating the quality of wood chips used in sulfate pulping according to the present invention. Figure 5 This is a flowchart of step S4 of a method for evaluating the quality of wood chips used in sulfate pulping according to the present invention. Detailed Implementation To better understand the technical content of this invention, a specific embodiment is provided below, and the invention will be further described in conjunction with the accompanying drawings.

[0024] See Figures 1 to 5 The present invention provides a method for evaluating the quality of wood chips used in sulfate pulping, comprising the following steps: Step S1: Collect a number of raw wood chips and divide them into multiple groups of test samples. After screening the test samples, obtain qualified wood chips and determine the qualified rate of the wood chips in the test samples. Step S2: Collect multidimensional thickness data of qualified wood chips through visual inspection, calculate adsorption uniformity, and calculate the uniformity score of the test sample based on adsorption uniformity. Step S3: Pulp qualified wood chips using the sulfate method, determine the pulp yield and fiber length, and calculate the pass rate score, yield score and length score of the test sample using the standardized score scoring method. Step S4: Based on the deep reinforcement learning algorithm, optimize the dynamic weight coefficients of uniformity score, pass rate score, yield score and length score; Step S5: Based on the dynamic weighting coefficient, the uniformity score, pass rate score, yield score and length score are weighted and summed to obtain the total quality score of the wood chips of the test sample. Step S6: Perform cumulative probability analysis based on the total quality score of all tested wood chips to classify the quality of raw wood chips into grades.

[0025] This invention discloses a method for evaluating the quality of wood chips used in sulfate pulping. This method is applicable to comparisons between wood chips from the same tree species. First, a certain quantity of raw wood chips from the same tree species is selected and divided into several groups of test samples of equal quantity. Each group of test samples is then scored using multi-dimensional evaluation indicators, including adsorption uniformity, wood chip qualification rate, pulp yield, and fiber length. Adsorption uniformity is calculated by considering factors such as wood chip thickness fluctuation and distribution skewness, reflecting the uniformity and consistency of alkali penetration during cooking, directly affecting canister blockage, screen residue, and pulp quality fluctuations. The wood chip qualification rate reflects the regularity of wood chip specifications, ensuring uniform chip size and avoiding excessive screen residue and uneven cooking. Pulp yield reflects the raw material conversion efficiency and economy, while fiber length determines the strength and applicable scenarios of the finished pulp. The scores of the four dimensions are weighted and summed to obtain the total wood chip quality score for each test sample. Then, a cumulative probability analysis is performed on the total wood chip quality scores of all test samples to determine the quality of the raw wood chips, providing data for raw material procurement.

[0026] After obtaining multiple groups of test samples, four scoring indicators need to be measured sequentially. The measurement has a strict order: first, each test sample is screened to identify qualified wood chips, and the qualified wood chip rate is determined based on these qualified wood chips and the raw wood chips in the test sample. Subsequent measurements of the other three scoring indicators must be based on qualified wood chips. First, multidimensional thickness data of qualified wood chips is collected through visual inspection. Based on this thickness data, adsorption uniformity can be calculated. After the adsorption uniformity calculation is completed, pulp yield and fiber length can be calculated simultaneously. For qualified wood chips in the same test sample, the sulfate pulping process is used, and the pulp yield is measured after cooking. By calculating the yield and fiber length, four scoring indicators can be obtained. Adsorption uniformity is scored separately, while wood chip pass rate, pulp yield, and fiber length are scored using a standardized scoring method. After determining the scores for the four scoring indicators, this invention introduces a deep reinforcement learning algorithm to optimize the dynamic weight coefficients of the four scoring indicators. This achieves optimal weights that match the real-time state of pulp production, significantly improving the accuracy, stability, and production adaptability of wood chip quality scoring. This enhances the accuracy of the final overall wood chip quality score calculation and addresses the economic benefits issues arising from raw material quality in the industrial production of wood chips.

[0027] Preferably, step S1 includes the following steps: Step S11: Purchase raw wood chips of the same tree species and divide them into multiple groups of test samples. Screen the test samples according to the TAPPI standard. Step S12: Screening to obtain raw wood chips with a diameter range of 9.5mm to 40mm as qualified wood chips; Step S13: Calculate the wood chip pass rate of the test sample based on the quality of qualified wood chips and raw wood chips.

[0028] The purchased raw wood chips of the same tree species are divided into several groups of test samples, such as 200 groups. The test samples are sieved in a sieve according to the TAPPI UM-21 standard method. The qualified wood chips are those with a diameter of 9.5mm to 40mm. Finally, the qualified wood chip pass rate of the test samples is calculated based on the quality of the qualified wood chips and the raw wood chips.

[0029] Preferably, step S2 includes the following specific steps: Step S21: Evenly distribute measurement points on the upper surface, lower surface, edge, and center of the qualified wood chips; Step S22: Use a convolutional neural network vision inspection system to obtain the original data of the wood chip thickness at the measurement points; Step S23: Based on the original wood chip thickness data, calculate the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution; Step S24: Calculate the adsorption uniformity U based on the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution, and obtain the uniformity score of the test sample using the inverse exponential scoring method. The formula for calculating the adsorption uniformity is:

[0030] The formula for calculating the uniformity score is:

[0031] in For uniform adsorption, For thickness standard deviation, The average thickness Due to extremely poor thickness, For thickness distribution skewness, A fixed correction factor, ranging from 0.92 to 1.05, is used to eliminate systematic errors in the testing equipment. This is the exponential decay coefficient, ranging from 1.5 to 2.5, used to enhance the scoring advantage of samples with excellent homogeneity. Let be the uniformity score of the i-th test sample. For the adsorption uniformity of the i-th test sample, The minimum adsorption uniformity among all tested samples, This represents the maximum adsorption uniformity across all tested samples.

[0032] After selecting qualified wood chips from the test samples, it is necessary to calculate the adsorption uniformity and uniformity score. Several measurement points are set on the qualified wood chips, and a convolutional neural network vision inspection system is used to acquire raw wood chip thickness data. The raw thickness data represents the measured thickness value at each measurement point. Based on this raw thickness data, the mean thickness, standard deviation thickness, range thickness, and thickness distribution skewness can be calculated. The mean thickness is the sum of the measured thickness values ​​at all measurement points divided by the number of measurement points. The standard deviation thickness is the sum of the squared deviations of all raw thickness data from the mean thickness. The range thickness is the maximum deviation among all raw thickness data. The difference between the thickness value and the minimum thickness value, and the thickness distribution skewness are calculated using Pearson's third-order moment skewness to reflect the symmetry of the thickness distribution. Finally, the adsorption uniformity is calculated based on the mean thickness, standard deviation of thickness, range of thickness, and thickness distribution skewness. Adsorption uniformity reflects whether the alkali solution penetration is uniform and consistent during the cooking process. After obtaining the adsorption uniformity, the inverse index scoring is used to determine the adsorption uniformity score. The design of the inverse index structure allows the score to increase faster as the adsorption uniformity is better, significantly amplifying the scoring advantage of high-quality samples, suppressing the score of poor-quality samples, making the grade distinction more obvious, and the scoring more in line with the actual pulping quality requirements.

[0033] Preferably, step S3, which involves pulping qualified wood chips using the sulfate process and determining the pulp yield and fiber length, includes the following specific steps: Step S31: After drying the qualified wood chips, put them into the digester and add cooking liquor to the digester. Set the corresponding cooking conditions to cook and pulp. Step S32: Wash the pulp obtained after cooking and calculate the pulp yield based on the sample's oven-dry weight. Step S33: Take the cooked and washed pulp and use a fiber analyzer to measure the fiber length.

[0034] Qualified wood chips after screening were selected as raw materials for pulping and cooking. The sulfate method was used to pulp the wood chips under the same cooking conditions. The pulp yield and fiber length after cooking were determined. The specific steps of pulping and cooking were as follows: qualified wood chips were naturally air-dried to a moisture content of about 7% to 15%. A certain amount of dried wood chips was weighed into a cooking tank, cooking liquor was added, and cooking and pulping were carried out under the appropriate cooking conditions. After cooking, the pulp obtained was washed, and the sample pulp was collected and weighed. The pulp yield was calculated based on the oven-dry weight of the sample. The cooking conditions for collecting laboratory sulfate pulping data are not unique. The laboratory cooking conditions should be close to the production conditions to be reflected so that the calculated wood chip score can be used to reflect the actual quality of the produced products. The conditions can be adjusted according to the production process to be evaluated. When scoring the same tree species, the data collected should be obtained under the same experimental cooking conditions.

[0035] The fiber length is the average length by weight of the fiber. The pulp after cooking and washing is measured using a fiber analyzer, and the test is carried out in accordance with the ISO 16065-2:2014 standard.

[0036] Preferably, step S3, which uses a standardized score scoring method to calculate the pass rate score, yield score, and length score, includes the following specific steps: Calculate the mean and standard deviation of the wood chip pass rate A, pulp yield Y, and fiber length F for the tested samples respectively; The standardized scoring method is used for scoring, and the scoring formula is as follows:

[0037]

[0038]

[0039] in These are the pass rate score, yield score, and length score for the i-th tested sample, respectively. These represent the wood chip pass rate, pulp yield, and fiber length of the i-th test sample, respectively. To detect the mean wood chip pass rate, mean pulp yield, and mean fiber length of the corresponding samples, The standard deviations of the wood chip pass rate, pulp yield, and fiber length corresponding to the test samples were determined.

[0040] After obtaining the wood chip pass rate, pulp yield, and fiber length, the mean and standard deviation are calculated uniformly. The mean of the wood chip pass rate is... and standard deviation The formula is as follows:

[0041]

[0042] Where n is the number of qualified wood chips, and the formulas for calculating the mean and standard deviation of pulp yield and fiber length are the same as those for calculating the wood chip qualification rate, and will not be repeated here. After determining the mean and standard deviation, a standardized score scoring method is used for scoring, which can eliminate the differences in the dimensions and numerical ranges of different scoring indicators, make the scores of each scoring indicator uniform and comparable, and ensure that the comprehensive evaluation results are objective, fair, stable and reliable.

[0043] Preferably, step S4 includes the following specific steps: Step S41: Construct the state space S of the pulping production environment. The state space is composed of the average wood chip qualification rate, the average pulp yield, the average fiber length, the average adsorption uniformity, the cooking energy consumption, the slag screening rate, and the tensile strength of the pulp. Step S42: Construct the action space M, which outputs 4-dimensional dynamic weight coefficients. Design a multi-objective reward function R, which comprehensively rewards improved slurry yield, achieved strength targets, reduced energy consumption, and improved adsorption uniformity. The expression for the multi-objective reward function is:

[0044] in The reward value for the current state action pair. This represents the actual slurry yield. To achieve the target slurry yield, For the actual tensile strength of the slurry, To achieve the target tensile strength of the slurry, For energy consumption, For maximum energy consumption, For the slag screening rate, To achieve the maximum slag screening rate, As a benchmark for adsorption uniformity, For the adsorption uniformity of the i-th test sample, For the weighting coefficients, satisfying .

[0045] Step S43: Using the DDPG dual network structure, the Actor network outputs actions based on the state space, and the Critic network evaluates the merits of the actions based on the reward function and performs feedback iteration. Step S44: Process the original weights output by the Actor network using softmax normalization constraints, and output the optimal dynamic weight coefficients. The dynamic normalization formula is as follows:

[0046] in Let be the final normalized dynamic weight for the k-th scoring indicator, k∈{A, U, Y, F}. This is the raw output of the Actor network. The weights from the previous time step are used to smooth out weight fluctuations. This is the penalty coefficient for weight fluctuations.

[0047] After determining the uniformity score, pass rate score, yield score, and length score, it is also necessary to determine the dynamic weight coefficients of the four-dimensional scoring indicators. This invention introduces a deep reinforcement learning algorithm. By constructing a state space, it can comprehensively perceive the multi-dimensional state information of pulp production and wood chip quality. By constructing an action space, it can accurately output the weight decision vector. The multi-objective reward function is designed to indicate the optimization direction of weight optimization that takes into account yield, strength, energy consumption, and uniformity. The Actor-Critic dual network structure adopted achieves adaptive optimization and accurate evaluation of weights through iteration. Finally, Softmax normalization with a penalty term ensures that the weights meet the constraints and fluctuate smoothly, so that the dynamic weights can be adaptively adjusted in real time to fit the actual production conditions, significantly improving the accuracy, stability, and adaptability of wood chip quality evaluation.

[0048] Preferably, the formula for calculating the total quality score of the wood chips is:

[0049] in The total score for the quality of wood chips in the i-th test sample. These represent the pass rate score, yield score, length score, and uniformity score for the i-th tested sample, respectively. These are dynamic weighting coefficients.

[0050] By calculating the total score for wood chip quality using a weighted average, different weights can be assigned according to the degree of influence of each scoring indicator on wood chip quality and pulping effect. On the basis of a unified dimension, multiple scoring indicators are organically integrated, so that the final score can not only reflect the merits and demerits of individual indicators, but also objectively reflect the overall quality of wood chips. The evaluation results are more in line with actual production needs and are more instructive.

[0051] Preferably, the grading rule in step S6 is as follows: Wood chips with a cumulative probability of ≤20% in the total quality score are classified as Grade 5 wood chips. Wood chips with a cumulative probability of 20% or less of the total quality score being ≤40% are classified as Grade 4 wood chips. Wood chips with a cumulative probability of 40% or less of the total quality score and ≤60% are classified as Grade 3 wood chips. Wood chips with a cumulative probability of 60% or less of the total quality score and ≤80% are classified as Grade II wood chips. A wood chip with a cumulative overall quality score of less than 80% is considered a first-class wood chip.

[0052] Finally, the overall quality score for the wood chips was determined. The Minitab software was used to perform cumulative probability analysis. Based on the scores corresponding to the 20% probability intervals of the distribution, five levels were defined to provide data support for raw material procurement.

[0053] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for evaluating the quality of wood chips used in sulfate pulping, characterized in that, Includes the following steps: Step S1: Collect a number of raw wood chips and divide them into multiple groups of test samples. After screening the test samples, obtain qualified wood chips and determine the qualified rate of the wood chips in the test samples. Step S2: Collect multidimensional thickness data of qualified wood chips through visual inspection, calculate adsorption uniformity, and calculate the uniformity score of the test sample based on adsorption uniformity. Step S3: Pulp qualified wood chips using the sulfate method, determine the pulp yield and fiber length, and calculate the pass rate score, yield score and length score of the test sample using the standardized score scoring method. Step S4: Based on the deep reinforcement learning algorithm, optimize the dynamic weight coefficients of uniformity score, pass rate score, yield score and length score; Step S5: Based on the dynamic weighting coefficient, the uniformity score, pass rate score, yield score and length score are weighted and summed to obtain the total quality score of the wood chips of the test sample. Step S6: Perform cumulative probability analysis based on the total quality score of all tested wood chips to classify the quality of raw wood chips into grades.

2. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The specific steps of step S1 include: Step S11: Purchase raw wood chips of the same tree species and divide them into multiple groups of test samples. Screen the test samples according to the TAPPI standard. Step S12: Screening to obtain raw wood chips with a diameter range of 9.5mm to 40mm as qualified wood chips; Step S13: Calculate the wood chip pass rate of the test sample based on the quality of qualified wood chips and raw wood chips.

3. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The specific steps of step S2 include: Step S21: Evenly distribute measurement points on the upper surface, lower surface, edge, and center of the qualified wood chips; Step S22: Use a convolutional neural network vision inspection system to obtain the original data of the wood chip thickness at the measurement points; Step S23: Based on the original wood chip thickness data, calculate the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution; Step S24: Calculate the adsorption uniformity based on the mean thickness, standard deviation of thickness, range of thickness, and skewness of thickness distribution, and obtain the uniformity score of the test sample using the inverse index scoring.

4. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 3, characterized in that, The formula for calculating the adsorption uniformity is: The formula for calculating the uniformity score is: in For uniform adsorption, For thickness standard deviation, The average thickness Due to extremely poor thickness, For thickness distribution skewness, To fix the correction factor, The exponential decay coefficient is... Let be the uniformity score of the i-th test sample. For the adsorption uniformity of the i-th test sample, The minimum adsorption uniformity among all tested samples, This represents the maximum adsorption uniformity across all tested samples.

5. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The specific steps in step S3, which involves pulping qualified wood chips using the sulfate process and determining the pulp yield and fiber length, include: Step S31: After drying the qualified wood chips, put them into the digester and add cooking liquor to the digester. Set the corresponding cooking conditions to cook and pulp. Step S32: Wash the pulp obtained after cooking and calculate the pulp yield based on the sample's oven-dry weight. Step S33: Take the cooked and washed pulp and use a fiber analyzer to measure the fiber length.

6. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The specific steps in step S3, which uses a standardized score scoring method to calculate the pass rate score, yield score, and length score, include: Calculate the mean and standard deviation of the wood chip pass rate A, pulp yield Y, and fiber length F for the tested samples respectively; The standardized scoring method is used for scoring, and the scoring formula is as follows: in These are the pass rate score, yield score, and length score for the i-th tested sample, respectively. These represent the wood chip pass rate, pulp yield, and fiber length of the i-th test sample, respectively. To detect the mean wood chip pass rate, mean pulp yield, and mean fiber length of the corresponding samples, The standard deviations of the wood chip pass rate, pulp yield, and fiber length corresponding to the test samples were determined.

7. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The specific steps of step S4 include: Step S41: Construct the state space S of the pulping production environment. The state space is composed of the average wood chip qualification rate, the average pulp yield, the average fiber length, the average adsorption uniformity, the cooking energy consumption, the slag screening rate, and the tensile strength of the pulp. Step S42: Construct an action space M, which outputs a 4-dimensional dynamic weight coefficient. Design a multi-objective reward function R to comprehensively reward the improvement of slurry yield, strength attainment, energy consumption reduction, and adsorption uniformity. Step S43: Using the DDPG dual network structure, the Actor network outputs actions based on the state space, and the Critic network evaluates the merits of the actions based on the reward function and performs feedback iteration. Step S44: Process the original weights output by the Actor network through softmax normalization constraints, and output the optimal dynamic weight coefficients.

8. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 7, characterized in that, The expression for the multi-objective reward function is: in The reward value for the current state action pair. This represents the actual slurry yield. To achieve the target slurry yield, For the actual tensile strength of the slurry, To achieve the target tensile strength of the slurry, For energy consumption, For maximum energy consumption, For the slag screening rate, To achieve the maximum slag screening rate, As a benchmark for adsorption uniformity, For the adsorption uniformity of the i-th test sample, These are the weighting coefficients.

9. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The formula for calculating the overall quality score of the wood chips is as follows: in The total score for the quality of wood chips in the i-th test sample. These represent the pass rate score, yield score, length score, and uniformity score for the i-th tested sample, respectively. These are dynamic weighting coefficients.

10. The method for evaluating the quality of wood chips used in sulfate pulping according to claim 1, characterized in that, The grading rules in step S6 are as follows: Wood chips with a cumulative probability of ≤20% in the total quality score are classified as Grade 5 wood chips. Wood chips with a cumulative probability of 20% or less of the total quality score being ≤40% are classified as Grade 4 wood chips. Wood chips with a cumulative probability of 40% or less of the total quality score and ≤60% are classified as Grade 3 wood chips. Wood chips with a cumulative probability of 60% or less of the total quality score and ≤80% are classified as Grade II wood chips. A wood chip with a cumulative overall quality score of less than 80% is considered a first-class wood chip.

Images