Adaptive control method for full-automatic washdown packaging scale unit
By using a fully automatic water-washing packaging scale unit and an improved sandcat optimization algorithm, the problems of adaptability and weighing error of the fully automatic packaging scale have been solved, achieving accurate weighing and efficient cleaning, thereby improving production efficiency and accuracy.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-09-25
- Publication Date
- 2026-04-02
AI Technical Summary
Existing fully automatic packaging scales lack adaptability during the weighing process and cannot flexibly adjust according to different material characteristics and packaging conditions, resulting in operational complexity and weighing errors. The cleaning process is also time-consuming and inconvenient.
The fully automatic water washing and packaging scale unit integrates water/air washing pipes and an arc-shaped flow guiding cleaning plate module. Combined with an improved sandcat optimization algorithm, including Piecewise chaotic mapping, Levy flight walk and sparrow warning mechanism, an adaptive control system is established to accurately control the material weight through fuzzy PID control.
The fully automatic water washing and packaging scale has achieved adaptive control, which improves weighing accuracy and cleaning efficiency, reduces operational complexity and weighing errors, and ensures production continuity and efficiency.
Smart Images
Figure CN2024120953_02042026_PF_FP_ABST
Abstract
Description
Self-adaptive control method of full-automatic water washing packaging scale unit TECHNICAL FIELD
[0001] The present application belongs to the technical field of intelligent control in the processing industry, and particularly relates to a self-adaptive control method of a full-automatic water washing packaging scale unit. BACKGROUND
[0002] The packaging scale is an important equipment widely used in the field of industrial production, mainly used for automatically completing the processes of weighing, metering and packaging in the commodity packaging process. With the popularization of large-scale and intelligent production, the full-automatic packaging scale plays a key role in improving production efficiency, reducing labor costs and ensuring product quality, and has high demand and broad market prospects.
[0003] Although the full-automatic scale has improved in terms of automation and accuracy, there are still the following problems: (1) the existing full-automatic scale lacks adaptability in the weighing process and cannot be flexibly adjusted according to the characteristics of different materials and packaging conditions, which leads to the need for manual setting of a series of parameters when processing different commodities, increasing the complexity and time cost of operation; (2) due to the influence of environmental factors, product shape, packaging materials and other factors in the actual use of the packaging scale, the weighing result may have certain errors; the existing full-automatic scale lacks real-time monitoring and correction mechanism for weighing errors, and cannot guarantee the accuracy and stability of the weighing result; (3) the traditional full-automatic scale has certain difficulties in the cleaning process, which requires more time and manpower; this limits its continuous use and efficiency on the production line; therefore, the present technical solution proposes a self-adaptive control method of a full-automatic water washing packaging scale unit.
[0004] SUMMARY
[0005] The present application provides a self-adaptive control method of a full-automatic water washing packaging scale unit, which solves the problems in the background art.
[0006] To solve the above technical problems, the present application is realized by the following technical solutions:
[0007] The application discloses a self-adaptive control method of a full-automatic water-washing packaging scale unit.
[0008] The self-adaptive control method of the full-automatic water-washing packaging scale unit comprises the following steps.
[0009] Step one: analyzing the full-automatic water-washing packaging scale, researching the action relationship between feeding and weight, and establishing a full-automatic water-washing packaging controller.
[0010] Step two: designing an improved sand cat optimization algorithm, introducing three improved points, introducing a Piecewise chaotic mapping theory to initialize the sand cat population in the population initialization stage, and introducing a Levy flight walking strategy in the attack stage.
[0011] Step three: after the sand cat hunting stage ends, a sparrow alert mechanism is introduced, when the sand cat group realizes danger, the edge sand cats quickly move to the safe area, and the sand cats in the population randomly walk to approach other sand cats, so that the convergence speed of the sand cat optimization algorithm is improved.
[0012] Step one is specifically:
[0013] According to the mechanical principle, a packaging scale mathematical model is established, since the impact force I(t) generated by the material falling at any time cannot be accurately detected, it is difficult to accurately predict the shutdown time t, therefore, an online adjustment strategy is adopted.
[0014] Assume that at the moment of shutdown t, the measured weight of the nth bag of material is M(n), the material falling weight is Ml(n), the impact force generated by the falling material is I(n), the actual weight of the material is S(n), the actual deviation at this time is e(n), and the nominal quantitative value is S, then: S(n) = M(n) + Ml(n) - I(n) (1) e(n) = S - S(n) (2) e(n) = S - M(n) + [I(n) - Ml(n)] (3)
[0015] Let X(n) = [I(n) - Ml(n)], X(n) represents the algebraic sum of the impact force and the weight of the falling material in the air when the nth bag of material is weighed, that is: e(n) = S - M(n) + X(n), n = 1, 2, 3, …, e(0) = 0 (4)
[0016] Obviously: M(n) = M0 + e(n-1) / 2 (5)
[0017] Where M0 is a fixed value, and 2 / 3S < M0 < S, e(n) = S - M0 - e(n-1) / 2 + X(n) (6) e(n-1) = S - M0 - e(n-2) / 2 + X(n-1) (7)
[0018] Subtracting equation (7) from equation (6) gives the iterative formula: e(n) - e(n-1) = 1 / 2[e(n-1) - e(n-2)] + [X(n) - X(n-1)] (8) e(n) - e(1) = 1 / 2[e(n-1) - e(0)] + [X(n) - X(1)] = -1 / 2e(n-1) + [X(n) - X(1)] (9) e(n) = X(n) - 1 / 2e(n-1) + e(1) - X(1) (10)
[0019] From equation (1), e(1) = S - M(1) + X(1), then: e(n) = -1 / 2e(n-1) + [X(n) + S - M(1)] (11)
[0020] From equation (2), M(1) = M0, then: e(n) = -1 / 2e(n-1) + [X(n) + S - M0] (12)
[0021] Through the above derivation, the current weight deviation of the material is obtained, and a mathematical model of the full-automatic water washing packaging scale is established, and a fuzzy PID method is used for control;
[0022] Step two is specifically:
[0023] The improved sand cat optimization algorithm is used to optimize the control parameters of PID, in which the basic sand cat optimization algorithm simulates the hunting behavior of sand cats, mainly including two aspects: search and predation, and the algorithm process is as follows:
[0024] First, the population is initialized: X i = l b +a×(u b -l b ) (13)
[0025] In the formula: X i represents the position of the i-th sand cat individual, a is a random number in [0,1], u b represents the upper limit, and l b represents the lower limit.
[0026] Hunting search stage:
[0027] Sand cats rely on emitting low-frequency noise signals to search for prey, and the signal sensing range is 2kHz to 0. With the iteration, the signal value r G is linearly reduced from 2 to 0, simulating the process that the sand cat gradually approaches the prey without losing or skipping, s M imitates the hearing characteristics of sand cats:
[0028] t is the current iteration number, T is the maximum iteration number, which can be flexibly set according to different problems to be solved, and the universality of the equation is shown.
[0029] The main parameter for controlling the transition of exploration and development process is R: R = 2 × r G × rand() (15)
[0030] In which, rand() represents a random number between 0 and 1, r G represents the process range of the sensitivity from 2 linearly decreasing to 0.
[0031] In the search step, the update of the current search position of each sand cat is based on a random position, which can make the sand cat explore new space in the search space. In order to prevent falling into local optimum, each sand cat has different sensitivity range r2: r2 = r G × rand() (16)
[0032] Therefore, r2 is used to update the position in the exploration and development stage, r G is used to guide the switching of parameter R in the two stages.
[0033] In the sand cat optimization algorithm, the sand cat updates its position based on the current population optimum position, its own position, and the sensitivity range. To ensure that the sand cat finds a new optimal position in the search area, which is between the current position and the target prey position, while ensuring the convergence effect and complexity of the algorithm, the following position update mathematical model is established:
[0034] Where x t best represents the position of the current sand cat population optimum individual.
[0035] Predation stage:
[0036] The sand cat determines the prey based on the signal heard by the ear, and the position update model in this stage is as follows:
[0037] Where rand() represents a random number between 0 and 1, the search range of the sand cat is considered as a circle, and the moving direction can be defined by a random angle θ on the circle, ranging from 0 to 360 degrees. Each sand cat in the population can move in a different circumferential direction to avoid falling into a local optimum; P is the distance of one step of the sand cat, x i t and x i t+1 are the positions of the i-th sand cat at time t and t+1, respectively.
[0038] Exploration and development of sand cat population:
[0039] The exploration and development of the sand cat population are guaranteed by the adaptive values of R and r G The sand cat algorithm performs development (hunting) when the R value is less than or equal to 1, and otherwise it can only perform exploration (searching):
[0040] Equation (19) represents the position update method of the i-th sand cat in the exploration and development stages. Different search methods for each sand cat avoid the algorithm falling into a local optimum and ensure the global search ability.
[0041] The Piecewise chaotic mapping is introduced to optimize the initialization of the sand cat population:
[0042] In the hunting process of the sand cat, the Levy flight strategy is introduced, which can provide a walking method with a step length conforming to the Levy distribution. However, in the predation stage, the distance between the sand cat and the prey is very close. In order to better conform to the hunting behavior of the sand cat, a regulation constant C = 0.37 is introduced to make the sand cat walk as close to the prey as possible. The Levy flight strategy formula is as follows:
[0043] Step three is:
[0044] After the end of the cat hunting stage, the sparrow alert mechanism is introduced to avoid the cat being discovered by its predators, thereby improving the convergence speed of the cat optimization algorithm.
[0045] When the cat realizes the danger, the cat at the edge of the group will quickly move to the safe area to obtain a better position, and the cat in the middle of the group will move randomly and approach other cats:
[0046] wherein x t best and x t worst respectively represent the current global best position and the worst position, beta is a step-controlled random number subject to [0,1] normal distribution, k represents the moving direction of the cat, which is a uniform random number in the range of [-1,1]. And respectively represent the current global best and worst fitness value, epsilon is a minimum constant to avoid the denominator being zero. When f i＝ f g indicates that the cat is at the edge of the group and is vulnerable to attacks by predators, and should move to a safe position as soon as possible.
[0047] The present application has the following beneficial effects relative to the prior art:
[0048] (1) The present scheme innovatively proposes a full-automatic water washing unit to overcome the defects of the prior art, such as the influence of material residues on the accuracy and the inconvenience of cleaning, and the water and gas dual cleaning is fast and residue-free. The present scheme also performs mechanical analysis on the feeding of the full-automatic water washing packaging scale, studies the influence of the discharging time on the weight of the material, and establishes a self-adaptive controller for the water washing packaging scale.
[0049] (2) The present scheme designs an improved cat optimization algorithm ISCSO, which introduces three improvements: introducing a Piecewise chaotic mapping strategy to initialize the cat population in the population initialization stage; introducing a Levy flight strategy to improve the global optimization ability of the algorithm in the population hunting stage, and introducing an adjustment constant to ensure the search speed; after the end of the hunting stage, the sparrow alert strategy is introduced to make the cat in the middle of the group approach the optimal position, and the cat in the worst position move away from the edge of the group, ensuring the search efficiency.
[0050] (3) Through the above optimization, the performance of the cat algorithm is effectively improved, the local optimum is avoided, and the control precision and self-adaptive adjustment ability of the full-automatic water washing packaging scale unit are realized. The experiment of the method provides a theoretical reference for the precise control of the weight in the production and packaging links, and has certain practical significance.
[0051] Of course, any product implementing the present application does not necessarily need to achieve all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS
[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the following embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.
[0053] Fig. 1 is a structural schematic diagram of a full-automatic water washing packaging scale unit of the present application;
[0054] Fig. 2 is an adaptive control system diagram corresponding to an adaptive control method of a full-automatic water washing packaging scale unit of the present application;
[0055] Fig. 3 is a flow step diagram of an adaptive control method of a full-automatic water washing packaging scale unit of the present application;
[0056] Fig. 4 is a convergence curve diagram of a first test function in an improved sand cat optimization algorithm and other multiple intelligent optimization algorithms in the embodiment of the present application;
[0057] Fig. 5 is a convergence curve diagram of a second test function in an improved sand cat optimization algorithm and other multiple intelligent optimization algorithms in the embodiment of the present application;
[0058] In the drawings, the component list represented by each reference numeral is as follows: 1, storage hopper; 2, water washing packaging scale shell; 3, load bearing unit; 4, quantitative packaging unit; 5, control unit; 6, water / gas washing pipeline. DETAILED DESCRIPTION
[0059] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.
[0060] As shown in Fig. 1, an adaptive control method of a full-automatic water washing packaging scale unit of the present application is realized based on ISCSO. The full-automatic water washing packaging scale unit includes a storage hopper 1, a full-automatic water washing unit, a load bearing unit 3, and a quantitative packaging unit 4. The full-automatic water washing unit is first integrated in the packaging scale and includes a water / gas intelligent spraying washing module 6 and an arc-shaped flow guide cleaning disc module.
[0061] The working control of the full-automatic water washing packaging scale includes the following steps:
[0062] Step one: analyzing the full-automatic water washing packaging scale discharging module, researching the action relationship between feeding and weight, and establishing a full-automatic water washing packaging controller;
[0063] Step two: design an improved sand cat optimization algorithm, introduce three improvements, in the population initialization stage, introduce the theory of Piecewise chaotic mapping, initialize the sand cat population, in the attack stage, introduce Levy flight strategy, because the distance between sand cat and prey is very close when attacking prey, in order to more conform to the behavior of sand cat attacking prey, introduce adjustment coefficient C = 0.37, make sand cat walk around prey as much as possible; The value of the adjustment coefficient is less than 1 to reduce the search time, which is an empirical value.
[0064] Step three: after the end of the sand cat hunting stage, introduce the sparrow alert mechanism, when the sand cat group realizes the danger, the edge sand cat moves to the safe area quickly, and the sand cat in the population randomly walks to other sand cats, so as to improve the convergence speed of the sand cat optimization algorithm
[0065] Step one is:
[0066] According to the principle of mechanics, a mathematical model of the packaging scale is established. Since it is difficult to accurately detect the size of the impact force I(t) generated by the material falling at any time, it is difficult to accurately predict the shutdown time t, therefore, an online adjustment strategy is adopted.
[0067] Assume that at the shutdown time t, the measured weight of the nth bag of material is M(n), the falling weight is Ml(n), the impact force generated by the falling material is I(n), the actual weight of the material is S(n), and the actual deviation is e(n) at this time, and the nominal quantitative value is S, then: S(n) = M(n) + Ml(n) - I(n) (1) e(n) = S - S(n) (2) e(n) = S - M(n) + [I(n) - Ml(n)] (3)
[0068] Let X(n) = [I(n) - Ml(n)], X(n) represents the algebraic sum of the impact force and the weight of the falling material when weighing the nth bag of material, that is: e(n) = S - M(n) + X(n), n = 1, 2, 3, …, e(0) = 0 (4
[0069] Obviously, M(n) = M0 + e(n-1) / 2 (5
[0070] Where M0 is a fixed value, and 2 / 3S < M0 < S, e(n) = S - M0 - e(n-1) / 2 + X(n) (6) e(n-1) = S - M0 - e(n-2) / 2 + X(n-1) (7)
[0071] Subtracting equation (7) from equation (6) gives the iterative formula: e(n)-e(n-1) = 1 / 2[e(n-1)-e(n-2)] + [X(n)-X(n-1)] (8) e(n)-e(1) = 1 / 2[e(n-1)-e(0)] + [X(n)-X(1)] = -1 / 2e(n-1) + [X(n)-X(1)] (9) e(n) = X(n)-1 / 2e(n-1) + e(1)-X(1) (10)
[0072] From equation (1), e(1) = S-M(1) + X(1), then: e(n) = -1 / 2e(n-1) + [X(n) + S-M(1)] (11)
[0073] From equation (2), M(1) = M0, then: e(n) = -1 / 2e(n-1) + [X(n) + S-M0] (12)
[0074] Through the above derivation, the current weight deviation of the material is obtained, and a mathematical model of the full-automatic washing and packaging scale is established, and a fuzzy PID mode is used for control.
[0075] The above formula and content are variables for deriving the mathematical model, but the final error expression does not contain these, so they do not have to be measured in actual application.
[0076] Step two is specifically:
[0077] The improved sand cat optimization algorithm optimizes the selection of the control parameters of the PID, wherein the basic sand cat optimization algorithm simulates the hunting behavior of the sand cat, mainly including two aspects: search and predation, and the algorithm process is as follows:
[0078] S1: first, population initialization: X i = l b +a×(u b -l b ) (13)
[0079] In the formula: X i represents the position of the i-th sand cat individual, a is a random number in [0, 1], u b represents the upper limit, and l b represents the lower limit; a is a variable for generating an initial population, so it is randomly selected between 0 and 1, and the upper limit and the lower limit are the limits of the search range.
[0080] S2: prey search stage:
[0081] The sand cat relies on emitting low-frequency noise signals to search for prey, and the signal sensing range is 2 kHz to 0. The intelligent optimization algorithm is a modeling algorithm that simulates animal habits, and the data is obtained from the published paper proposing the algorithm. As the iteration proceeds, the signal value r G Linearly decreases from 2 to 0, simulating the process of the sand cat gradually approaching the prey without losing or skipping, s M The hearing characteristics of the sand cat are imitated:
[0082] t is the current iteration number, T is the maximum iteration number, and this value can be flexibly set according to different problems to be solved, showing the universality of the equation.
[0083] The main parameter controlling the transition of the exploration and development process is R: R = 2 x r G x rand() (15)
[0084] where rand() represents a random number between 0 and 1, r G represents the process of linearly decreasing sensitivity from 2 to 0.
[0085] In the search step, the update of the current search position of each sand cat is based on a random position, which can make the sand cat explore new space in the search space. In order to prevent falling into local optimum, each sand cat has a different sensitivity range r2: r2 = r G x rand() (16)
[0086] Therefore, r2 is used to update the position in the exploration and development stages, r G is used to guide the switching of parameter R in the two stages.
[0087] In the sand cat optimization algorithm, the sand cat updates its position based on the optimal position of the current population, its own position, and the sensitivity range. In order to ensure that the sand cat finds a new optimal position in the search area, which is located between the current position and the target prey position, while ensuring the convergence effect and complexity of the algorithm, the following position update mathematical model is established:
[0088] where x t best represents the position of the optimal individual in the current sand cat population.
[0089] S3: Predation stage:
[0090] The sand cat determines the prey according to the signal heard by the ear, and the position update model in this stage is as follows:
[0091] where rand() represents a random number between 0 and 1, the search range of the cat is considered as a circle, and the moving direction of the cat can be defined by a random angle θ on the circle, ranging from 0 to 360 degrees. Each cat in the population can move along a different circumferential direction to avoid falling into a local optimum. P is the distance of one step of the cat, and x i t and x i t+1 are the positions of the i-th cat at time t and t+1, respectively.
[0092] Exploration and development of the cat population:
[0093] The exploration and development of the cat population are guaranteed by the adaptive values of R and r G When the value of R is less than or equal to 1, the cat algorithm performs development (hunting), otherwise it can only perform exploration (searching):
[0094] Equation (19) represents the position update method of the i-th cat in the exploration and development stages. Different search methods for each cat avoid the algorithm falling into a local optimum and ensure global search ability.
[0095] Improvement point 1:
[0096] In step two, the Piecewise chaotic mapping is introduced to optimize the initialization of the cat population:
[0097] Improvement point 2:
[0098] In step two, the Levy flight strategy is introduced during the cat hunting process, which can provide a walking method with a step length conforming to the Levy distribution. However, during the hunting stage, the distance between the cat and the prey is very close. In order to better conform to the hunting behavior of the cat, a regulation constant C = 0.37 is introduced to make the cat walk as close to the prey as possible. The Levy flight strategy formula is as follows:
[0099] Levy is a strategy for updating data, which is relatively complex;
[0100] Improvement point 3:
[0101] In step two, after the cat hunting stage, the sparrow alert mechanism is introduced to avoid the cat being discovered by its predator, thereby improving the convergence speed of the cat optimization algorithm.
[0102] When the cat realizes the danger, the cat at the edge of the group will quickly move to the safe area to obtain a better position, and the cat in the middle of the group will move randomly and approach other cats:
[0103] where x t best and x t worst represent the current global best position and the worst position, respectively. β is a step- controlled random number obeying the normal distribution [0, 1], k represents the moving direction of the sand cat, which is a uniform random number in the range [-1, 1]. and represent the current global best and worst fitness values, respectively, and ε is a minimum constant to avoid the denominator being zero. i and f g represent the fitness value of the current population and the global best fitness value, respectively, and f i = f g indicates that the sand cat is at the edge of the population and is vulnerable to predators, so it should move to a safe position as soon as possible.
[0104] In Figure 2, the PID controller is optimized by the improved sand cat optimization algorithm. Through the mathematical model established for the fully automatic washing packaging scale, precise control of the discharging module of the packaging scale is achieved, ensuring that the deviation between the expected weight and the actual weight of the packaging is controlled within a reasonable range.
[0105] An ISCSO-based fully automatic washing packaging scale unit and its adaptive control system, which uses a basic test function to verify the performance of the improved algorithm. The test function is selected from CEC2020, which is an authoritative test set of test functions, containing about 20 functions to test algorithm performance. The improved sand cat optimization algorithm is compared with the basic sand cat algorithm, particle swarm optimization algorithm, Harris eagle optimization algorithm, artificial gorilla troop optimization algorithm, sparrow search algorithm, and hippocampus optimization algorithm to verify the performance of the improved sand cat optimization algorithm. To ensure fairness, the population size of each algorithm is set to 50, and the maximum number of iterations is set to 500.
[0106] As shown in Figure 4, the improved sand cat algorithm has a greater advantage in convergence speed and optimization ability compared to other algorithms.
[0107] An ISCSO-based fully automatic washing packaging scale unit and its adaptive control system, which uses a basic test function to verify the performance of the improved algorithm. The test function is selected from CEC2020, which is an authoritative test set of test functions, containing about 20 functions to test algorithm performance. The improved sand cat optimization algorithm is compared with the basic sand cat algorithm, particle swarm optimization algorithm, Harris eagle optimization algorithm, artificial gorilla troop optimization algorithm, sparrow search algorithm, and hippocampus optimization algorithm to verify the performance of the improved sand cat optimization algorithm. To ensure fairness, the population size of each algorithm is set to 50, and the maximum number of iterations is set to 500.
[0108] As shown in Figure 5, it can be seen that the improved sand cat algorithm has greater advantages in convergence speed and optimization ability compared with other algorithms.
[0109] The Convergence curve in Figure 4 and Figure 5 is a convergence curve, Dim is a dimension number, Best score F* is a test function name, the vertical coordinate axis represents a function value, and the horizontal coordinate axis is an iteration number.
[0110] Working principle: Taking the full-automatic water washing packaging scale unit as the research object, the material is put into the material storage hopper by the feeding device, the load receiving device receives and weighs the material quality, the quality is returned to the feeding device for material quality adjustment, the quantitative packaging device encapsulates the material package meeting the quality requirements, and the full-automatic water washing device cleans the inside of the packaging scale through water and gas spraying, and prepares for the next weighing and packaging.
[0111] The preferred embodiments disclosed above are only used to help explain the present application. The preferred embodiments do not describe all the details and limit the present application to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the present application. The present application is selected and described in detail to better explain the principles and practical application of the present application, so that those skilled in the art can well understand and utilize the present application. The present application is limited by the claims and their full scope and equivalents.
Claims
1. A method of adaptive control of a fully automatic wash-pack scale unit, the adaptive control method applied to a wash-pack scale unit comprising: The water washing packaging scale shell (2), the storage hopper (1) installed on the upper part of the water washing packaging scale shell (2), the full-automatic water washing unit, the bearing unit (3) and the quantitative packaging unit (4); the full-automatic water washing unit is integrated in the water washing packaging scale shell (2), and the full-automatic water washing unit comprises a water / gas washing pipeline (6) and an arc-shaped guide flow cleaning disc module; characterized by comprising the following steps: Step one, analyze the full-automatic water washing packaging scale discharge module, study the relationship between feeding and weight, and establish a full-automatic water washing packaging controller; According to the mechanical principle, a packaging scale mathematical model is established. Since it is difficult to accurately detect the size of the impact force I(t) generated by the material falling at any time, it is difficult to accurately predict the shutdown time t, therefore, an online adjustment strategy is adopted; Assume that at the shutdown time t, the measured weight of the nth bag of material is M(n), the material falling weight is Ml(n), the impact force generated by the material falling is I(n), the actual weight of the material is S(n), the actual deviation at this time is e(n), and the nominal quantitative value is S, then: S(n)=M(n)+Ml(n)-I(n) (1) e(n)=S-S(n) (2) e(n)=S-M(n)+[I(n)-Ml(n)] (3) Let X(n)=[I(n)-Ml(n)], X(n) represents the algebraic sum of the impact force and the weight of the falling material when the nth bag of material is weighed, that is: e(n)=S-M(n)+X(n), n=1,2,3,..., e(0)=0 (4) Obviously: M(n)=M0+e(n-1) / 2 (5) Where M0 is a fixed value, and 2 / 3S e(n)=S-M0-e(n-1) / 2+X(n) (6) e(n-1)=S-M0-e(n-2) / 2+X(n-1) (7) Subtracting equation (7) from equation (6) gives the iterative formula: e(n)-e(n-1)=1 / 2[e(n-1)-e(n-2)]+[X(n)-X(n-1)] (8) e(n)-e(1)=1 / 2[e(n-1)-e(0)]+[X(n)-X(1)]=-1 / 2e(n-1)+[X(n)-X(1)] (9) e(n)=X(n)-1 / 2e(n-1)+e(1)-X(1) (10) From equation (1), e(1)=S-M(1)+X(1), then: e(n)=-1 / 2e(n-1)+[X(n)+S-M(1)] (11) From equation (2), M(1)=M0, then: e(n)=-1 / 2e(n-1)+[X(n)+S-M0] (12) Through the above derivation, the current weight deviation of the material is obtained, the mathematical model of the full-automatic water washing packaging scale is established, and the fuzzy PID mode is adopted for control; Step two, design an improved sand cat optimization algorithm, and introduce three improvement points; In the population initialization stage, the theory of Piecewise chaotic mapping is introduced to initialize the cat population; in the attack stage, the Levy flight strategy is introduced; because the distance between the cat and the prey is very close when attacking the prey, in order to better conform to the behavior of the cat attacking the prey, the adjustment coefficient C=0.37 is introduced, so that the cat walks around the prey; Specifically: The improved cat optimization algorithm optimizes the selection of the control parameters of PID, wherein the basic cat optimization algorithm simulates the hunting behavior of the cat, mainly including two aspects: search and predation, and the algorithm process is as follows: First, the population is initialized: X i = l b + a x (u b - l b ) (13) wherein: X i represents the position of the ith individual sand cat, a is a random number in [0,1], u b represents the upper bound, l b represents the lower bound; The prey search stage: The sand cat relies on emitting low-frequency noise signals to search for prey, the signal perception range is 2 kHz to 0, as the iteration proceeds, the signal value r G Linearly decreases from 2 to 0, simulates the process that the sand cat gradually approaches the prey and does not lose or skip, s M The hearing characteristics of the sand cat are imitated: t is the current iteration number, T is the maximum iteration number, which is flexibly set according to different problems to be solved, and the universality of the equation is shown; The main parameter for controlling the transition of the exploration and development process is R: R = 2 x r G x rand() (15) where rand() represents a random number between 0 and 1, r G represents the range of the process in which the sensitivity linearly decreases from 2 to 0; In the search step, the update of the current search position of each cat is based on a random position, which enables the cat to explore new space in the search space, and in order to prevent falling into local optimum, each cat has a different sensitivity range r2: r2 = r G xrand() (16) Thus, r2 is used for position updating in the exploration and development phase, r G is used to guide the switching of the parameter R in the two phases; In the cat optimization algorithm, the cat will update its position according to the optimal position of the current population, its own position and the sensitivity range r2: Sensitivity range update position, in order to ensure that the sand cat finds a new optimal position in the search area, the position is located between the current position and the target prey position, while ensuring the convergence effect and complexity of the algorithm, the following position update mathematical model is established: where x t best represent the position of the best individual of the current group of sand cats. The predation stage: The sand cat discriminates the prey according to the signal heard by the ear, and the position updating model in this stage is as follows: where rand() represents a random number between [0, 1], the search range of sand cat is considered as a circle, and the moving direction of sand cat is defined by a random angle θ on the circle, ranging from 0 to 360 degrees. Each sand cat in the population moves along a different circumferential direction to avoid falling into local optimum; P is the distance of one step of sand cat, x i t and x i t+1 are the positions of the i-th sand cat at time t and t+1, respectively. Exploration and development of the cat population: Exploration and exploitation of the sand cat population is governed by the adaptive values of R and rG. The sand cat algorithm exploits (hunts) when the value of R is less than or equal to 1, otherwise it can only explore (search): Equation (19) represents the position update method of the i-th cat in the exploration and development stage, and the different search methods of each cat avoid the algorithm falling into local optimum and ensure the global search ability; The introduction of Piecewise chaotic mapping is used to optimize the initialization of sand cat population: Levy flight strategy is introduced in the hunting process of sand cat, and a walking method with step length conforming to Levy distribution is provided; in the predation stage, the distance between sand cat and prey is very close. In order to be more consistent with the predation behavior of sand cat, a regulation constant C=0.37 is introduced to make the sand cat walk around the prey, and the Levy flight strategy formula is as follows: Step three, after the cat hunting stage, the sparrow alert mechanism is introduced, when the cat group realizes the danger, the edge cat moves quickly to the safe area, the cat in the population randomly walks and approaches other cats, avoids being discovered by its predators, and thus improves the convergence speed of the cat optimization algorithm; When the cat realizes the danger, the cat on the edge of the group will quickly move to the safe area to get a better position, and the cat in the middle of the group will move randomly and approach other cats: where x t best and x t worst respectively represent the current global best position and the worst position; β is a step- controllable random number subject to the normal distribution [0, 1], k represents the direction of the sand cat, which is a uniform random number in the range [-1, 1]; and respectively represent the current global best and worst fitness values, ε is the minimum constant to avoid the denominator being 0; when f i = f g , it indicates that the sand cat is at the edge of the population and is easily affected by the local optimum. The attack of the predator, at this time, it should be moved to a safe position as soon as possible.