An AI algorithm-based pulp concentration intelligent production line regulation method and system

Abstract

The application discloses a pulp concentration intelligent production line regulation and control method and system based on an AI algorithm, and relates to the technical field of pulp concentration control. The method comprises the following steps: establishing a feeding model to obtain the pulp adding amount and the water adding amount under a test pulp concentration and when the standard loading volume is v; obtaining the test pulp concentration comprises the following steps: a plurality of connecting pipes are connected at each height in the stirring barrel, the outlet of the connecting pipe is fixedly communicated with the top of the stirring barrel, and a detector and a liquid pump are connected on the pipe body of the connecting pipe. The application is provided with a feeding model, the target concentration can be determined according to the quality requirements of papermaking and the use requirements of users on paper, the target concentration is input into the pulp adding amount model and the water adding amount model, the target pulp adding amount and the target water adding amount are obtained respectively, and after mixing is completed, the mixed solution with the standard loading volume under the target concentration is obtained. The pulp batching system does not need to be customized in advance.

Description

Technical Field

[0001] This invention relates to the field of pulp concentration control technology, specifically to a method and system for intelligent production line control of pulp concentration based on AI algorithms. Background Technology

[0002] The main processes of making tissue paper include pulping, refining, mixing, forming, pressing, drying, creping and winding.

[0003] In the pulping process, pulp concentration directly affects the paper forming quality. Too low a concentration leads to fiber agglomeration, increased fiber mobility, and a longer time required to form a flocculated state, thus affecting the paper's uniformity and overall quality. Conversely, too high a concentration may increase the pulp's viscosity and fluidity, making it difficult for the pulp to flow in pipes, even causing blockages. It also hinders the uniform dispersion of fibers. Furthermore, pulp concentration determines the paper's mechanical strength; both too low and too high concentrations weaken the paper's toughness, making it prone to breakage. Real-time measurement and control of pulp concentration prevents drastic fluctuations during papermaking, ensuring the stability and continuity of the production process. This helps improve production efficiency and reduce costs. Precise concentration control also contributes to consistent and stable product quality, meeting customer needs and expectations.

[0004] Chinese invention patent application, publication number CN106283806A, discloses a method and system for controlling pulp quality in a high-consistency refining system. This method is based on the input and output data of the high-consistency disc refiner measured by sensors on the chemimechanical pulping production line. It determines the order of the sub-model by combining the AIC criterion and obtains the sub-model parameters by using the forgetting factor least squares method. Combined with the mechanism model of the pulp quality freeness index, a Wiener model structure of the high-consistency refining system is established. The quadratic performance index is optimized by using a sequential quadratic programming algorithm to achieve effective control of the pulp quality index - freeness - output from the pulping process.

[0005] Existing technologies and existing pulp batching systems still have certain shortcomings in practical applications, such as:

[0006] Based on the quality of papermaking and the different needs of users for paper, customers need to adjust the pulp concentration during production. Existing pulp batching systems are generally customized in advance according to user needs, which lacks sufficient flexibility and cannot quickly adapt to the different requirements of different customers or different products for pulp concentration.

[0007] After adjusting the pulp concentration, separate pulp preparation is required. Paper production can only proceed after the pulp preparation is completed, resulting in poor production continuity and reduced production efficiency.

[0008] In the pulp preparation process, the pulp mixture is usually stirred and mixed by the agitator in the mixing tank. However, the agitator consumes a lot of power when stirring the pulp mixture, which is not conducive to energy saving of the system and increases production costs. Summary of the Invention

[0009] The purpose of this invention is to provide a method and system for intelligent control of pulp concentration in a production line based on AI algorithms, so as to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution: a method for intelligent control of pulp concentration in a production line based on AI algorithms, comprising:

[0011] Establish a feeding model to obtain the amount of pulp and water added when the test pulp concentration is and the standard container volume is v.

[0012] Obtaining the test pulp concentration involves the following steps:

[0013] Several connecting pipes are connected at each height inside the mixing tank. The outlet of the connecting pipe is fixedly connected to the top of the mixing tank. A detector and a liquid pump are connected to the body of the connecting pipe.

[0014] Water at different heights in the mixing tank is pumped to the corresponding connecting pipes by a liquid pump. The pulp concentration is detected by a detector to obtain the concentration sets at different heights: [(a011, a012, ..., a01N), (a021, a022, ..., a02N), ..., (a0N1, a0N2, ..., a0NN)];

[0015] Calculate the average value of all detected concentrations on the same horizontal plane to obtain the detected concentration dataset (a01, a02, ..., a0N);

[0016] Based on different detection heights, the data in the detection concentration dataset are weighted and averaged to obtain the test pulp concentration c, c=e1×a01+e2×a02+...+eN×a0N;

[0017] Based on the feeding model, a pulp addition model and a water addition model are obtained. The target concentration is input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively.

[0018] Establish a stirring model to obtain the stirring time model of the stirrer. Input the target concentration into the stirring time model of the stirrer to obtain the target stirring time of the stirrer in the mixing tank.

[0019] Establish a refining model, input the target concentration into the refining model, and obtain the target refining time of the refining machine.

[0020] Furthermore, the establishment of the feeding model includes the following steps:

[0021] Weigh out several portions of pulp, with the weights of the pulp being m1, m2, ..., mN, to obtain the set of pulp weights;

[0022] Obtain the standard container volume of the mixing tank in the production line, where the standard container volume is v;

[0023] Weigh out a number of portions of water with a volume of v that are equal in weight to the pulp.

[0024] Water was added to the mixing tank separately, and then the pulp in the pulp weight set was added to the mixing tank separately. After mixing thoroughly, the total volume of pulp and water was measured as v1, v2, ..., vN, and the test pulp concentration in the mixing tank was measured as ρ1, ρ2, ..., ρN. The total volume set of pulp and water and the test pulp concentration set were obtained respectively.

[0025] Subtract the total volume of the obtained pulp and water from the standard holding volume of the mixing tank to obtain the set of volume increases, which are: v1-v, v2-v, ..., vN-v;

[0026] Divide the data in the set of volume increase by the data in the set of total volume of pulp and water respectively to obtain the set of volume change ratios, which are: (v1-v) / v1, (v2-v) / v2, ..., (vN-v) / vN;

[0027] Multiply the data in the set of pulp weights by the data in the set of volume change ratios to obtain the set of pulp reduction amounts, which are: m1×(v1-v) / v1, m2×(v2-v) / v2, ..., mN×(vN-v) / vN. Then, obtain the set of pulp addition amounts, which are: m1×[1-(v1-v) / v1], m2×[1-(v2-v) / v2], ..., mN×[1-(vN-v) / vN].

[0028] Take out the mixture with volume increases of v1-v, v2-v, ..., vN-v respectively, and obtain the volume of water in the mixture by evaporation. This gives the set of water reduction volumes, which are p1, p2, ..., pN respectively. Then, the water addition volumes are v-p1, v-p2, ..., v-pN. Finally, the water addition amounts are ρwater×(v-p1), ρwater×(v-p2), ..., ρwater×(v-pN).

[0029] Furthermore, a linear fit is performed on the set of test pulp concentrations and the set of pulp addition amounts to obtain a pulp addition amount model;

[0030] A linear fit was performed on the set of test pulp concentrations and the set of water addition amounts to obtain the water addition amount model.

[0031] Furthermore, the establishment of the stirring model includes the following steps:

[0032] Obtain the number of storage tanks n, the storage volume q of the storage tanks, and the discharge time t of each storage tank, where (number of storage tanks n-1) × storage volume q of the storage tanks = standard loading volume v of the mixing tank;

[0033] The residence time of water and pulp in the mixing tank is calculated as T = (number of storage tanks n-1) × discharge time t of each storage tank;

[0034] The concentrations of water and pulp in the mixing tank were measured to obtain the mixing time of the agitator in the mixing tank at the test pulp concentration.

[0035] Furthermore, the method for determining the stirring time of the agitator in the mixing tank to achieve the aforementioned test pulp concentration is as follows:

[0036] The stirring time jx of the agitator when the pulp and water mixture in the mixing tank reaches the test pulp concentration is obtained, and its set is [j1, j2, ..., jN];

[0037] The set of test pulp concentrations was linearly fitted to the set of agitator agitation times to obtain the agitator agitation time model.

[0038] Furthermore, the method for establishing the pulping model of the pulping machine is as follows:

[0039] Subtract the residence time T of water and pulp in the mixing tank from the mixing time jx of the agitator at the test pulp concentration to obtain the set of gap times. Calculate the average value of the gap times to obtain the average gap time X.

[0040] Obtain the power consumption P1 per unit time of the pulper and the power consumption P2 per unit time of the agitator, and calculate P1 / P2=k (k<1).

[0041] If the average interval time X > 2 × the stirring time jx of the agitator under the test pulp concentration, the target refining time tx of the refiner is jx / (k×2).

[0042] If the stirring time jx of the agitator under the test pulp concentration is less than or equal to the average interval time X, and is less than or equal to 2 × the stirring time jx of the agitator under the test pulp concentration, then the target stirring time tx of the agitator is k × P2 × jx / (P1 × 2.5).

[0043] If the average interval time X < the stirring time jx of the agitator under the test pulp concentration, the target refining time tx of the refiner is k × P2 × jx / (P1 × 3).

[0044] Furthermore, the method also includes:

[0045] Determine the time b for the mixing tank to replenish the storage tank, and calculate the residence time T of water and pulp in the mixing tank = (number of storage tanks n-1) × discharge time t of each storage tank - mixing time b.

[0046] A pulp concentration intelligent production line control system based on AI algorithms, using a pulp concentration intelligent production line control method, includes:

[0047] The data acquisition module acquires data on pulp concentration detected by each detector in the mixing tank, the number of storage tanks n, the storage volume q of each storage tank, the discharge time t of each storage tank, and the time b for the mixing tank to replenish the storage tanks.

[0048] The calculation module establishes a feeding model to obtain the amount of pulp and water added when the test pulp concentration is met and the standard container volume is v. Based on the feeding model, it obtains a pulp addition model and a water addition model. The target concentration is input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively. A stirring model is established to obtain the stirring time model of the stirrer. The target concentration is input into the stirring time model of the stirrer to obtain the target stirring time of the stirrer. A refining model is established. The target concentration is input into the refining model to obtain the target refining time of the refiner.

[0049] The AI ​​algorithm module, based on deep learning and training of the model in the computing module, enables intelligent control of pulp concentration in the production line.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] This AI-based intelligent production line control method and system for pulp concentration features a feeding model that determines the target concentration based on papermaking quality requirements and user needs for paper applications. The target concentration is then input into the pulp addition model and water addition model to obtain the target pulp addition amount and target water addition amount, respectively. After mixing, a mixture with a standard container volume at the target concentration is obtained. This pulp batching system does not require pre-customization and can quickly adapt to different customers or different products' requirements for pulp concentration, offering good flexibility.

[0052] Meanwhile, the concentration of the mixture gradually increases during the stirring process. In this solution, the pulp and water quantities are controlled based on the standard container volume, thereby achieving the target concentration under the conditions of maximum water and minimum pulp quantity, effectively saving pulp and reducing production costs.

[0053] In addition, multiple storage tanks are provided, which sequentially feed materials to the Yankee drying cylinder, thus achieving uninterrupted material supply and increasing production efficiency. Furthermore, if the target concentration needs to be changed during operation, the previous target concentration of the mixed liquor in the storage tank is used up, and the materials are mixed according to the changed target concentration within the residence time T of water and pulp in the mixing tank. This ensures good production continuity of the pulp mixing system and improves production efficiency.

[0054] In addition, a pulping model is provided, which can adjust the pulping time of the refiner according to the residence time T of water and pulp in the mixing tank and the interval time. This reduces the mixing time of the agitator while ensuring the pulp concentration, thereby reducing the energy consumption of the system. Attached Figure Description

[0055] Figure 1 This is a flowchart of the present invention;

[0056] Figure 2 This is a front view of the distribution of the pulp concentration tester of the present invention;

[0057] Figure 3 This is a top view of the distribution of the pulp concentration testing instrument of the present invention;

[0058] Figure 4 This is a schematic diagram showing the connection between the mixing tank and the storage tank of the present invention;

[0059] Figure 5 This is a power consumption diagram of the stirrer and grinder of the present invention;

[0060] Figure 6 The graph shows the change in pulp concentration in the mixing tank when the agitator of this invention is working alone and when the agitator and the grinder are working simultaneously.

[0061] In the diagram: 1. Detector; 2. Liquid pump; 3. Water inlet pipe; 4. Pulp inlet pipe; 5. Electrically controlled valve; 6. Yankee drying cylinder. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] like Figures 1-6 As shown, the present invention provides a technical solution: a method for intelligent control of pulp concentration in a production line based on AI algorithms, comprising:

[0064] A feeding model was established to obtain the amount of pulp and water added at the test pulp concentration and with a standard container volume of v. The establishment of the feeding model includes the following steps:

[0065] Weigh out several portions of pulp. In this scheme, the pulp is produced by breaking up the pulp board with a pulper. The particle size control range of the broken pulp includes 65% -2μm content and 100% -45μm content. The weights of the pulp are m1, m2, ..., mN, respectively, to obtain the pulp weight set (m1, m2, ..., mN).

[0066] Obtain the standard container volume of the mixing tank in the production line. The standard container volume refers to the maximum amount of solution that the mixing tank can mix in a single batch. The standard container volume is v. Weigh out several portions of water with a volume v equal to the amount of pulp and add the water to the mixing tank. Then, add the pulp from the pulp weight set to the mixing tank. After mixing thoroughly, measure the total volume of pulp and water as v1, v2, ..., vN, and measure the test pulp concentration in the mixing tank as ρ1, ρ2, ..., ρN. Obtain the total volume set of pulp and water (v1, v2, ..., vN) and the test pulp concentration set (ρ1, ρ2, ..., ρN). In this scheme, thorough mixing means that the concentration of the mixture of pulp and water in the mixing tank reaches the corresponding test pulp concentration.

[0067] like Figure 2 and Figure 3 As shown, obtaining the test pulp concentration includes the following steps:

[0068] Several connecting pipes are connected at each height inside the mixing tank, and the outlet of the connecting pipe is fixedly connected to the top of the mixing tank. A detector 1 is connected to the body of each connecting pipe. A liquid pump 2 is connected to the main pipe. When the liquid pump 2 is started, the liquid in the mixing tank flows in the connecting pipe. The pulp concentration is detected by the detector 1. The detector can be a spectrometer or other device that can test the pulp concentration of the mixed liquor.

[0069] like Figure 4As shown, the water inlet pipe 3 is located at the bottom of the mixing tank to inject water into it. The pulper is located at the top of the mixing tank and is connected to it via the pulp inlet pipe 4. Thus, when the agitator inside the mixing tank stirs the solution, the concentration of the solution at the top of the tank will be lower than that at the bottom. Therefore, in this design, three sets of connecting pipes are installed at different heights. The distance between the topmost connecting pipe set and the liquid surface is 1 / 5 of the total liquid surface height; the distance between the middle connecting pipe set and the liquid surface is 3 / 5 of the total liquid surface height; and the distance between the bottommost connecting pipe set and the liquid surface is 4 / 5 of the total liquid surface height. To ensure accuracy, each set of connecting pipes includes three connecting pipes for detecting the mixture at different positions at the same height.

[0070] Water at different heights in the mixing tank is pumped into corresponding connecting pipes using a liquid pump. The pulp concentration is then measured using a detector, resulting in concentration sets at different heights: [(a011, a012, ..., a01N), (a021, a022, ..., a02N), ..., (a0N1, a0N2, ..., a0NN)]. The average value of all detected concentrations at the same horizontal plane is calculated to obtain the detection concentration dataset (a01, a02, ..., a0N). Finally, based on different detection heights, the data in the detection concentration dataset are weighted and averaged to obtain the test pulp concentration c, where c = e1×a01 + e2×a02 + ... + eN×a0N. In this scheme, since there are three sets of connecting pipes in the height direction, c = e1×a01 + e2×a02 + e3×a03.

[0071] like Figure 4 As shown, in this scheme, the mixed solution enters the storage tank through corresponding pipes, and then is sequentially transported to the surface of the Yankee drying cylinder through the storage tank and a series of precisely controlled conveying devices. During the liquid flow, the pulp concentration in the liquid gradually increases, thereby reaching the test pulp concentration. In this way, it is not necessary to directly stir the pulp concentration to the test pulp concentration in the mixing tank, reducing system energy consumption. In addition, since the concentration value in the middle of the mixing tank best represents the overall pulp concentration (the concentration is higher towards the bottom), in this scheme, the weights e1, e2, and e3 are set to 0.15, 0.6, and 0.25, respectively. Assuming the test pulp concentration is 30%, when the weighted average pulp concentration obtained by the agitator in the mixing tank is 20%, the agitator in the mixing tank will continue to stir the mixture. When the weighted average pulp concentration is 18%, the concentration deviation value is 10%, and this concentration deviation value of 10% can be compensated for in subsequent liquid conveying (verified by those skilled in the art through a limited number of experiments). At this time, the controller controls the agitator to stop working.

[0072] Subtract the total volume of pulp and water from the standard capacity of the mixing tank to obtain a set of volume increases, specifically: v1-v, v2-v, ..., vN-v. Divide the data in the set of volume increases by the data in the set of total volume of pulp and water to obtain a set of volume change ratios, specifically: (v1-v) / v1, (v2-v) / v2, ..., (vN-v) / vN.

[0073] Multiply the data in the set of pulp weights with the data in the set of volume change ratios to obtain the set of pulp reduction amounts, which are: m1×(v1-v) / v1, m2×(v2-v) / v2, ..., mN×(vN-v) / vN. Then, obtain the set of pulp addition amounts, which are: m1×[1-(v1-v) / v1], m2×[1-(v2-v) / v2], ..., mN×[1-(vN-v) / vN].

[0074] Take out the mixture with volume increases of v1-v, v2-v, ..., vN-v respectively, and obtain the volume of water in the mixture by evaporation. This gives the set of water reduction volumes, which are p1, p2, ..., pN respectively. Then, the water addition volumes are v-p1, v-p2, ..., v-pN. Finally, the water addition amounts are ρwater×(v-p1), ρwater×(v-p2), ..., ρwater×(v-pN).

[0075] Finally, the set of test pulp concentrations and the set of pulp addition amounts are linearly fitted to obtain the pulp addition amount model, and the set of test pulp concentrations and the set of water addition amounts are linearly fitted to obtain the water addition amount model.

[0076] Before production, the target concentration is determined based on the paper quality requirements and the user's needs for the paper's intended use. The target concentration is then input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively. For example, if the user sets the pulp concentration to 30%, the target concentration of 30% is input into the pulp addition model and the water addition model, resulting in a pulp amount of 25kg and a water amount of 100kg. The water enters the mixing tank directly through the water inlet pipe 3. After the pulp is ground to the set particle size, it enters the mixing tank through the pulp inlet pipe 4. After mixing, a mixture with a standard container volume at the target concentration is obtained.

[0077] In addition, it is understandable that the concentration of the mixture gradually increases during the stirring process. However, in this solution, the pulp and water quantities are controlled based on the standard container volume, thereby achieving the target concentration under the conditions of maximum water and minimum pulp quantity, effectively saving pulp and reducing production costs.

[0078] A stirring model is established to obtain the stirring time model of the stirrer. The target concentration is input into the stirring time model of the stirrer to obtain the target stirring time of the stirrer in the mixing tank.

[0079] The establishment of the stirring model includes the following steps:

[0080] Obtain the number of storage tanks n, the storage volume q of each tank, and the discharge time t of each tank. Here, (number of tanks n-1) × storage volume q = standard loading volume v of the mixing tank. Calculate the residence time T of water and pulp in the mixing tank = (number of tanks n-1) × discharge time t of each tank. Determine the concentrations of water and pulp in the mixing tank to obtain the mixing time required to reach the test pulp concentration. It can be known that this mixing time is the time required for both water and pulp to be added, and for the agitator to stir the mixture, thus achieving the required mixing time for the corresponding test pulp concentration.

[0081] In addition, to improve the accuracy of the system, the method also includes determining the machine time b for the mixing tank to replenish the storage tank. The machine time b includes the material pump start-up time and the feeding time. The residence time T of water and pulp in the mixing tank is calculated as T = (number of storage tanks n-1) × discharge time t of each storage tank - machine time b.

[0082] In one specific embodiment of this solution, four storage tanks are provided, and each storage tank is connected to an electrically controlled valve 5 on both sides of the pipe. Initially, all four storage tanks are full of pulp. During operation, the electrically controlled valves 5 are opened sequentially from top to bottom. When three storage tanks are empty, the electrically controlled valve on the left side of the empty storage tank is opened and the storage valve on the right side of the empty storage tank is closed to simultaneously replenish the three storage tanks. After the three storage tanks are replenished, the mixing tank is empty. At this time, water and pulp are added to the mixing tank according to the target concentration.

[0083] This system is equipped with multiple storage tanks, which sequentially feed materials into the Yankee drying cylinder, thus achieving uninterrupted material supply and increasing production efficiency. In addition, if the target concentration needs to be changed during operation, the previous target concentration of the mixed liquor in the storage tank is used up, and the materials are mixed according to the changed target concentration within the residence time T of water and pulp in the mixing tank. This ensures good production continuity of the pulp mixing system and improves production efficiency.

[0084] The mixing time of the agitator will vary depending on the pulp concentration. In this scheme, the method for determining the mixing time of the agitator in the mixing tank to achieve the test pulp concentration is as follows:

[0085] The stirring time jx of the agitator when the pulp and water mixture in the mixing tank reaches the test pulp concentration is obtained, and its set is [j1, j2, ..., jN]. The set of test pulp concentrations is linearly fitted with the set of agitator stirring times to obtain the target stirring time of the agitator in the mixing tank at the target concentration.

[0086] like Figure 5 and Figure 6 As shown, it is understandable that the more dispersed and smaller the particle size of the pulp, the shorter the mixing time under the same mixing conditions. Figure 6 ② in the figure represents the time-concentration curve of the mixture after the pulp has been ground for a certain period of time and then stirred by the agitator. ① represents the time-concentration curve of the mixture after stirring by the agitator at the initial pulp particle size. The power consumption per unit time of the agitator is greater than that of the refiner. Considering the correlation between the residence time T of water and pulp in the mixing tank and the stirring time of the agitator, in this scheme, a refining model is established, and the target concentration is input into the refining model to obtain the target refining time of the refiner.

[0087] Specifically, the method for establishing the pulping model of the pulping machine is as follows:

[0088] Subtract the residence time T of water and pulp in the mixing tank from the mixing time jx of the agitator at the test pulp concentration to obtain the set of gap times. Considering that the long gap time has little impact on the whole system, the average value of the gap time is calculated to obtain the average gap time X for the convenience of data processing later. The power consumption P1 per unit time of the refiner and the power consumption P2 per unit time of the agitator are obtained. P1 / P2 = k is calculated (k < 1).

[0089] If the average interval time X > 2 × the agitation time jx of the agitator at the test pulp concentration, the target refining time tx of the refiner is tx = jx / (k × 2). If the agitation time jx of the agitator at the test pulp concentration is ≤ average interval time X ≤ 2 × the agitation time jx of the agitator at the test pulp concentration, the target agitation time tx of the agitator is tx = jx / (k × 2.5). If the average interval time X < the agitation time jx of the agitator at the test pulp concentration, the target refining time tx of the refiner is tx = jx / (k × 3).

[0090] The coefficients in the denominator of the formula for the target refining time tx of the refiner are different in the three cases. This is because it takes into account the conveying time of the pulp particles, the time it takes to fall from the pulp inlet pipe to the water surface, and the maximum dissolution time of the pulp after it falls to the water surface. This is to prevent insufficient mixing of the pulp while minimizing the agitator mixing time.

[0091] This solution also discloses an AI-based intelligent production line control system for pulp concentration, employing an intelligent production line control method for pulp concentration, including:

[0092] The data acquisition module acquires data on pulp concentration detected by each detector in the mixing tank, the number of storage tanks n, the storage volume q of each storage tank, the discharge time t of each storage tank, and the time b for the mixing tank to replenish the storage tanks.

[0093] The calculation module establishes a feeding model to obtain the amount of pulp and water added when the test pulp concentration is met and the standard container volume is v. Based on the feeding model, it obtains a pulp addition model and a water addition model. The target concentration is input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively. A stirring model is established to obtain the stirring time model of the stirrer. The target concentration is input into the stirring time model of the stirrer to obtain the target stirring time of the stirrer. A refining model is established. The target concentration is input into the refining model to obtain the target refining time of the refiner.

[0094] The AI ​​algorithm module, based on deep learning and training of the model in the computing module, enables intelligent control of pulp concentration in the production line.

[0095] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended embodiments and their equivalents.

Claims

1. A method for intelligent control of pulp concentration in a production line based on AI algorithms, characterized in that, include: Establish a feeding model to obtain the amount of pulp and water added when the test pulp concentration is and the standard container volume is v. Obtaining the test pulp concentration involves the following steps: Several connecting pipes are connected at each height inside the mixing tank. The outlet of the connecting pipe is fixedly connected to the top of the mixing tank. A detector and a liquid pump are connected to the body of the connecting pipe. Water at different heights in the mixing tank is pumped to the corresponding connecting pipes by a liquid pump. The pulp concentration is detected by a detector to obtain the concentration sets at different heights: [(a011, a012, ..., a01N), (a021, a022, ..., a02N), ..., (a0N1, a0N2, ..., a0NN)]; Calculate the average value of all detected concentrations on the same horizontal plane to obtain the detected concentration dataset (a01, a02, ..., a0N); Based on different detection heights, the data in the detection concentration dataset are weighted and averaged to obtain the test pulp concentration c, c=e1×a01+e2×a02+...+eN×a0N; Based on the feeding model, a pulp addition model and a water addition model are obtained. The target concentration is input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively. Establish a stirring model to obtain the stirring time model of the stirrer. Input the target concentration into the stirring time model of the stirrer to obtain the target stirring time of the stirrer in the mixing tank. Establish a refining model, input the target concentration into the refining model, and obtain the target refining time of the refining machine; The establishment of the feeding model includes the following steps: Weigh out several portions of pulp, with the weights of the pulp being m1, m2, ..., mN, to obtain the set of pulp weights; Obtain the standard container volume of the mixing tank in the production line, where the standard container volume is v; Weigh out a number of portions of water with a volume of v that are equal in weight to the pulp. Water was added to the mixing tank separately, and then the pulp in the pulp weight set was added to the mixing tank separately. After mixing thoroughly, the total volume of pulp and water was measured as v1, v2, ..., vN, and the test pulp concentration in the mixing tank was measured as ρ1, ρ2, ..., ρN. The total volume set of pulp and water and the test pulp concentration set were obtained respectively. Subtract the total volume of the obtained pulp and water from the standard holding volume of the mixing tank to obtain the set of volume increases, which are: v1-v, v2-v, ..., vN-v; Divide the data in the set of volume increase by the data in the set of total volume of pulp and water respectively to obtain the set of volume change ratios, which are: (v1-v) / v1, (v2-v) / v2, ..., (vN-v) / vN; Multiply the data in the set of pulp weights by the data in the set of volume change ratios to obtain the set of pulp reduction amounts, which are: m1×(v1-v) / v1, m2×(v2-v) / v2, ..., mN×(vN-v) / vN. Then, obtain the set of pulp addition amounts, which are: m1×[1-(v1-v) / v1], m2×[1-(v2-v) / v2], ..., mN×[1-(vN-v) / vN]. Take out the mixture with volume increases of v1-v, v2-v, ..., vN-v respectively, and obtain the volume of water in the mixture by evaporation. This gives the set of water reduction volumes, which are p1, p2, ..., pN respectively. Then, the water addition volumes are v-p1, v-p2, ..., v-pN. Finally, the water addition amounts are ρwater×(v-p1), ρwater×(v-p2), ..., ρwater×(v-pN).

2. The method for intelligent control of pulp concentration in a production line based on AI algorithm according to claim 1, characterized in that, The set of test pulp concentrations and the set of pulp addition amounts are linearly fitted to obtain the pulp addition amount model; A linear fit was performed on the set of test pulp concentrations and the set of water addition amounts to obtain the water addition amount model.

3. The method for intelligent control of pulp concentration in a production line based on AI algorithm according to claim 1, characterized in that, The establishment of the stirring model includes the following steps: Obtain the number of storage tanks n, the storage volume q of the storage tanks, and the discharge time t of each storage tank, where (number of storage tanks n-1) × storage volume q of the storage tanks = standard loading volume v of the mixing tank; The residence time of water and pulp in the mixing tank is calculated as T = (number of storage tanks n-1) × discharge time t of each storage tank; The concentrations of water and pulp in the mixing tank were measured to obtain the mixing time of the agitator in the mixing tank at the test pulp concentration.

4. The method for intelligent control of pulp concentration in a production line based on AI algorithm according to claim 2, characterized in that, To achieve the required test pulp concentration, the method for determining the stirring time of the agitator in the mixing tank is as follows: The stirring time jx of the agitator when the pulp and water mixture in the mixing tank reaches the test pulp concentration is obtained, and its set is [j1, j2, ..., jN]; The set of test pulp concentrations was linearly fitted to the set of agitator agitation times to obtain the agitator agitation time model.

5. The method for intelligent control of pulp concentration in a production line based on AI algorithm according to claim 4, characterized in that, The method for establishing the pulping model of the pulping machine is as follows: Subtract the residence time T of water and pulp in the mixing tank from the mixing time jx of the agitator at the test pulp concentration to obtain the set of gap times. Calculate the average value of the gap times to obtain the average gap time X. Obtain the power consumption P1 per unit time of the pulper and the power consumption P2 per unit time of the agitator, and calculate P1 / P2=k, k<1; If the average interval time X > 2 × the stirring time jx of the agitator under the test pulp concentration, the target refining time tx of the refiner is jx / (k×2). If the stirring time jx of the agitator under the test pulp concentration is less than or equal to the average interval time X, and is less than or equal to 2 × the stirring time jx of the agitator under the test pulp concentration, then the target stirring time tx of the agitator is k × P2 × jx / (P1 × 2.5). If the average interval time X < the stirring time jx of the agitator under the test pulp concentration, the target refining time tx of the refiner is k × P2 × jx / (P1 × 3).

6. The method for intelligent control of pulp concentration in a production line based on AI algorithm according to claim 3, characterized in that, The method further includes: Determine the time b for the mixing tank to replenish the storage tank, and calculate the residence time T of water and pulp in the mixing tank = (number of storage tanks n-1) × discharge time t of each storage tank - mixing time b.

7. A pulp concentration intelligent production line control system based on AI algorithm, using the pulp concentration intelligent production line control method according to any one of claims 1-6, characterized in that, include: The data acquisition module acquires data on pulp concentration detected by each detector in the mixing tank, the number of storage tanks n, the storage volume q of each storage tank, the discharge time t of each storage tank, and the time b for the mixing tank to replenish the storage tanks. The calculation module establishes a feeding model to obtain the amount of pulp and water added when the test pulp concentration is 100% and the standard container volume is v. Based on the feeding model, it obtains a pulp addition model and a water addition model. The target concentration is input into the pulp addition model and the water addition model to obtain the target pulp addition amount and the target water addition amount, respectively. The mixing model is established to obtain the mixing time model of the mixer. The target concentration is input into the mixing time model of the mixer to obtain the target mixing time of the mixer tank. The refining model is established to obtain the target refining time of the refiner. The AI ​​algorithm module, based on deep learning and training of the model in the computing module, enables intelligent control of pulp concentration in the production line.

Images