Method for regulating and controlling quantity of liquid-phase discharge plasma sparks

Abstract

The invention discloses a liquid-phase discharge plasma spark quantity regulation and control method which specifically comprises the following steps: step 1, collecting a discharge spark image to obtain the spark quantity in the discharge spark image; 2, the spark number in the discharge process and the set spark number are solved, and the deviation and the deviation change rate of the spark number are obtained; 3, introducing a dynamic telescopic factor to adjust a discourse domain on the basis of the spark quantity deviation and the change rate; step 4, performing fuzzification processing on the input and output variable expansion and contraction factors and establishing a membership function; 5, designing a fuzzy rule for the dynamic telescopic factor; step 6, performing fuzzy reasoning according to the membership degree of the telescopic factor signal and a fuzzy rule, and performing defuzzification; and 7, regulating and controlling the number of discharge sparks according to the output reference voltage obtained by defuzzification. The problem that the spark number cannot be responded and regulated accurately in real time due to fluctuation of the discharge spark number in the prior art is solved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology for industrial processes, and relates to a method for controlling the number of sparks in liquid phase discharge plasma. Background Technology

[0002] Liquid-phase discharge plasma is a technology that utilizes plasma discharge in a liquid-phase environment to process materials, and spark discharge occurs during the process. Plasma electrolytic oxidation (PEO) is a surface modification method belonging to liquid-phase discharge plasma technology. Light metals (such as aluminum, magnesium, and titanium) and their alloys are subjected to plasma discharge under liquid-phase conditions to generate a dense ceramic layer mainly composed of oxides on the surface of the workpiece. This ceramic layer has good corrosion resistance, wear resistance, and insulation properties.

[0003] In the PEO process, the number of sparks is a core characteristic parameter reflecting the quality and discharge state of the film. Its evolution determines the final ceramic layer thickness and density, which are evaluation indicators of film performance. The PEO control system uses an optimized spark number characteristic curve as a closed-loop process control variable. The system adjusts the reaction process in real time according to the plasma reaction state, thereby avoiding the influence of changes in other process parameters and achieving real-time and precise response and control of the PEO process. Operators often rely on their experience in observing the patterns of discharge phenomena as a basis for auxiliary process control. However, relying solely on manual experience cannot effectively quantify the number of sparks, let alone accurately control the process in real time.

[0004] Liquid phase discharge (PEO) processes are characterized by nonlinearity and time-varying properties, with the number of sparks exhibiting complex dynamic changes across reaction stages. Therefore, nonlinear intelligent control methods adapted to these complex characteristics are required. Fuzzy control methods, with their strong robustness to parameter disturbances and external interference, are widely used in nonlinear intelligent control. However, due to the multiple reaction stages in the PEO discharge process and the wide-ranging dynamic changes in the number of sparks, deviations or rates of change in the number of sparks can exceed preset limits. Traditional fuzzy control methods cannot provide real-time, precise response and control over the number of sparks. Summary of the Invention

[0005] This invention addresses a method for controlling the number of sparks during liquid phase discharge, solving the problem in existing technologies where fluctuations in the number of discharge sparks prevent real-time, accurate response and control of the spark quantity.

[0006] The technical solution adopted in this invention is a method for controlling the number of sparks in liquid phase discharge plasma, which specifically includes the following steps: Step 1: Acquire an image of the discharge spark and obtain the number of sparks in the image; Step 2: Calculate the deviation and rate of change of the number of sparks obtained from the set number of sparks during the discharge process. Step 3: Based on the spark quantity deviation and rate of change, a dynamic scaling factor is introduced to adjust the domain of discourse; Step 4: Fuzzyize the scaling factors of the input and output variables and establish membership functions; Step 5: Design fuzzy rules for the dynamic scaling factor; Step 6: Perform fuzzy inference based on the membership degree and fuzzy rules of the scaling factor signal, and then perform defuzzification; Step 7: Adjust the number of discharge sparks based on the output reference voltage obtained from the defuzzification.

[0007] The invention is further characterized by: The specific process of step 2 is as follows: Subtract the number of discharge sparks during the liquid phase discharge process from the set target number of sparks. ,get t Spark count deviation at any moment e ( t and the corresponding rate of change of deviation ec ( t ), (1) (2) in, e ( t () represents the deviation in the number of sparks at time t. e ( t -1) represents the spark count deviation at time t-1.

[0008] The specific process of step 3 is as follows: The scaling factor of the input variable for constant discharge spark quantity control is The scaling factor of the output variable is The basic universe of discourse of the scaled input variables X ( e )and X ( ec and the fundamental universe of discourse of the output variable Y(u) They can be represented as follows: (3) in, For deviation e Boundary values ​​of the fundamental universe of discourse Rate of change of deviation ec Boundary values ​​of the fundamental universe of discourse and The scaling factor for the input variables. and For output variables u The lower and upper bounds of the fundamental domain, This is the scaling factor for the output variable.

[0009] The specific process of step 4 is as follows: Step 4.1, define the quantization universe of the scaling factors of the two input variables. X =[ x l , x h ], x l and x h Given the lower and upper bounds of the quantization universe of the input scaling factor, output the quantization universe of the input scaling factor. Y =[ y l , y h ], y l and y h The quantization domain of the output variable scaling factor, and the error in the number of discharge sparks. e Quantification factor K e Error change rate ec Quantification factor K ec and output reference voltage u Quantification factor K u The calculation formula is: (4) (5) (6) in, and The quantization domain is [0.5, 2.0], and the corresponding fuzzy set is {S, MS, M, MB, B}; The quantization domain is [0.5, 1.0], and the corresponding fuzzy set is {S, MS, Z}; Step 4.2, and The trapezoidal membership function consists of four parameters. a , b , c , d Definition, where a < b < c < d ,but and Membership function Represented as: (7) The triangular membership function consists of three parameters. a , b , c Defined, then Membership function Represented as: (8) in, a < b < c .

[0010] Step 5 involves designing the input and output variables of the fuzzy controller. Input variables e and ec The six linguistic variables {NB, NM, NS, PS, PM, PB} are used to describe the values, which correspond to negative large, negative medium, negative small, positive small, positive medium, and positive large, respectively. e The range is between [-500, 500]. ec The range is between [-250, 250], and the range of language variables is divided as follows: when e The range is between [-500, -333]. ec When the range is between [-250, -167], it is classified as NB; when e The range is between [-333, -167]. ec When the range is between [-167, -83], it is classified as NM; when e The range is between [-167, 0]. ec When the range is between [-83, 0], it is divided into NS; when e The range is between [0, 167]. ec When the range is between [0, 83], it is classified as PS; when e The range is between [167, 333]. ec When the range is between [83, 167], it is classified as PM; when e The range is between [333, 500]. ec When the range is between [167, 250], it is classified as PB; By introducing a scaling factor and performing fuzzing processing. and The domain range is quantized to [0.5, 2.0], and the membership function is defined as follows: NB: Coverage range is [0.5, 0.8], core range is [0.6, 0.7], decreasing to both sides; NM: Coverage range is [0.7, 1.0], core interval is [0.8, 0.9], decreasing on both sides; NS: Coverage range is [0.9, 1.2], core interval is [1.0, 1.1], decreasing on both sides; PS: The coverage area is [1.1, 1.4], the core interval is [1.2, 1.3], and the range decreases on both sides; PM: Coverage range is [1.3, 1.6], core range is [1.4, 1.5], decreasing on both sides; PB: Coverage range is [1.6, 2.0], core interval is [1.7, 1.9], decreasing on both sides; Each fuzziness level The membership degree is calculated as follows: NB: (9) NM: (10) NS: (11) PS: (12) PM: (13) PB: (14) The domain range is quantized to [0.5, 1.0], and the membership function is defined as follows: NB: Center point 0.55, coverage area [0.5, 0.6], decreasing on both sides; NM: Center point 0.65, coverage area [0.58, 0.72], decreasing on both sides; NS: Center point 0.75, coverage area [0.7, 0.8], decreasing on both sides; PS: Center point 0.85, coverage area [0.78, 0.92], decreasing on both sides; PM: Center point 0.95, coverage area [0.9, 1.0], decreasing on both sides; PB: Center point 0.99, coverage area [0.98, 1.0], decreasing on both sides; Each fuzziness level The membership degree is calculated as follows: NB: (15) NM: (16) NS: (17) PS: (18) PM: (19) PB: (20) Among them, output variables and The language values ​​are described using B, MB, M, MS, and S; output variables The linguistic variables are described by Z, MS, and S.

[0011] The specific process of step 6 is as follows: Step 6.1: Based on the fuzzy rule base and the fuzzified input signal, fuzzy logic operations are used to derive the fuzzy control output; Step 6.2: Use the centroid method to defuzzify the fuzzy control output.

[0012] The specific process of step 6.1 is as follows: Step 6.1.1, determine the current... e and ec The fuzzy membership degree is evaluated to determine whether it meets the rule preconditions and to calculate the applicability of the activation rule. Specifically: Rule prerequisites: If a certain rule's " e The corresponding fuzzy set e Membership degree > 0, and " ec The corresponding fuzzy set ec If the membership degree is greater than 0, the rule is activated and incorporated into subsequent reasoning; If the membership degree of any input variable to the fuzzy set of the rule premises is equal to 0, then the rule does not match and is excluded from the reasoning process.

[0013] For each activation rule, the applicability is calculated using the minimum value operation, with the following formula: (twenty one) in, This represents the applicability of a single activation rule, with a value range of [0,1]. , Represents the membership degree of a fuzzy subset with a variable universe of discourse; Step 6.1.2, for the three output variables of each activation rule , , Using applicability respectively The upper bound of the fuzzy set of conclusions is truncated using the following formula: (twenty two) in, The original membership function of the fuzzy set of the output variable conclusion. , The trapezoidal membership function, Let the membership function of the triangle be... This is the truncated membership function, retaining only the membership values ​​≤ applicability from the original membership function. The portion exceeding the limit is removed to determine the matching rule's fit. The maximum value of the fuzzy sets of conclusions after truncating all activation rules is used to synthesize a unified fuzzy control output set for the output variable, as shown in the formula: (twenty three) in, This represents the set of synthesized fuzzy control outputs.

[0014] In step 6.2, the formula for defuzzification is as follows: (twenty four) in, express The first in i The fuzzy control output values ​​corresponding to each fuzzy subset. This represents the scaling factor of the output universe of discourse. This represents the output value after defuzzification.

[0015] The beneficial effect of this invention is that, addressing the problems of insufficient adaptive capability and low precision in controlling the number of discharge sparks due to manual experience-based judgment or offline experimental parameter optimization, this invention proposes a method for controlling the number of discharge sparks in liquid-phase discharge plasma. By introducing a scaling factor and adjusting the domain of discourse in real time, it can adapt to the time-varying nature of the number of sparks at different reaction stages during liquid-phase discharge, accurately control the number of discharge sparks, better adapt to process parameters, and make the prepared ceramic layer more stable. Attached Figure Description

[0016] Figure 1 This is a block diagram illustrating the principle of variable universe of discourse fuzzy control (VUFC) in the liquid phase discharge plasma spark quantity control method of the present invention. Figure 2 The spark count is extracted by an image processing algorithm in the liquid phase discharge plasma spark count control method of the present invention. Figures 3(a) to 3(c) show the input variables in the liquid phase discharge plasma spark quantity control method of the present invention. e and ec Membership function and output variable u Membership function; Figures 4(a) to 4(b) are the target spark curve, feedback curve under VUFC, and output reference voltage diagrams in the liquid phase discharge plasma spark quantity control method of the present invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0018] Example 1 The present invention provides a method for controlling the number of sparks in liquid-phase discharge plasma, such as... Figure 1 As shown, it includes the following steps: Step 1: Real-time acquisition of discharge spark images using an industrial camera. First, image enhancement is performed using grayscale conversion and histogram peak-based grayscale stretching. Then, weighted image segmentation improves the segmentation accuracy of the spark region. Next, the watershed algorithm preserves the integrity of spark point boundaries and further extracts spark boundary information. Finally, connected component labeling is used to extract the spark quantity features. The extracted spark quantity feature map is shown below. Figure 2 ; Step 2, use the set target number of sparks. Subtract the number of discharge sparks during the liquid phase discharge process ,get t Spark count deviation at any moment e ( t and the corresponding rate of change of deviation ec ( t ), (1) (2) in, e ( t () represents the deviation in the number of sparks at time t. e ( t -1) represents the spark count deviation at time t-1.

[0019] Step 3: Based on the spark quantity deviation and rate of change, a dynamic scaling factor is introduced to adjust the domain of discourse; Step 4: Fuzzyize the scaling factors of the input and output variables and establish membership functions; Step 5: Design fuzzy rules for the dynamic scaling factor; Step 6: Perform fuzzy inference based on the membership degree and fuzzy rules of the scaling factor signal, and then perform defuzzification; Step 7: The output reference voltage signal obtained from the defuzzification is input to the PEO power supply. The PEO power supply adjusts the number of sparks by applying the output reference voltage to the load.

[0020] Example 2 Step 3 involves the following steps: Addressing the spark quantity deviation... e and the rate of change of deviation ec Based on this, a scaling factor is introduced to dynamically adjust the universe of discourse. Fuzzy inference is performed according to the membership function and fuzzy rules to defuzzify the fuzzy control output and obtain the output voltage. u The number of discharge sparks in the system is accurately controlled. In this embodiment, the deviation... e The fundamental universe of discourse is [-500, 500], and the rate of change of the deviation is... ec The fundamental universe of discourse is [-250, 250], and the output reference voltage is... u The fundamental universe of discourse is [0, 40]. Specifically, for a dual-input single-output control system, the scaling factor of the input variable of the VUFC controller for discharging spark quantity control is set to... The scaling factor of the output variable is This enables dynamic and adaptive adjustment of the universe of discourse size for the input and output variables of the fuzzy controller. The basic universe of discourse for the expanded input variables... X ( e )and X ( ec and the fundamental universe of discourse of the output variable Y(u) They can be represented as follows: (3) in For deviation e Boundary values ​​of the fundamental universe of discourse Rate of change of deviation ec Boundary values ​​of the fundamental universe of discourse and The scaling factor for the input variables. and For output variables u The lower and upper bounds of the fundamental domain, This is the scaling factor for the output variable.

[0021] Example 3 The specific process of step 4 is as follows: Step 4.1, fuzzify the scaling factors of the input and output variables. Define the quantization domain of the scaling factors of the two input variables. X =[ x l ,x h ], x l and x h Given the lower and upper bounds of the quantization universe of the input scaling factor, output the quantization universe of the input scaling factor. Y =[ y l , y h ], y l and y h The quantization universe is the scaling factor for the output variable. Mapping from the fundamental universe to the quantization universe requires the quantization factor. K To achieve the error in the number of discharge sparks e Quantification factor K e Error change rate ec Quantification factor K ec and output reference voltage u Quantification factor K u The calculation formula is: (4) (5) (6) and The quantization domain is [0.5, 2.0], and the corresponding fuzzy set is {S, MS, M, MB, B}; The quantization domain is [0.5, 1.0], and the corresponding fuzzy set is {S, MS, Z}.

[0022] Step 4.2: Establish membership functions for the input and output variables, and convert their fuzzification results into corresponding linguistic variables. and Choosing a trapezoidal membership function can reduce the system's sensitivity to instantaneous errors, making the control more stable. Choosing the triangular membership function can improve the system's response speed. and The trapezoidal membership function consists of four parameters. a , b , c , d Definition, where a < b < c < d ,but and Membership function It can be represented as: (7) The triangular membership function consists of three parameters. a , b , c Definition, where a < b < c ,but Membership function It can be represented as: (8) Based on the quantization domain range of the set input and output variables, the input variables of the fuzzy controller for the number of discharge sparks are... e and ec The membership functions are shown in Figures 3(a) and 3(b), and the output variables are... u The membership function is shown in Figure 3(c).

[0023] Example 4 Step 5 is specifically as follows: based on the input discharge spark quantity error e and its error change rate ec Scaling factor for the universe of discourse , and Design fuzzy rules to form a fuzzy controller with two inputs and three outputs.

[0024] Design the input and output variables of the fuzzy controller: Input variables e and ec The language uses six variable values ​​{NB, NM, NS, PS, PM, PB} to describe the values, which correspond to negative large, negative medium, negative small, positive small, positive medium, and positive large, respectively.

[0025] Specifically, e The range is between [-500, 500]. ec The range is between [-250, 250], and the range of language variables is divided as follows: when e The range is between [-500, -333]. ec When the range is between [-250, -167], it is classified as NB; when e The range is between [-333, -167]. ec When the range is between [-167, -83], it is classified as NM; when eThe range is between [-167, 0]. ec When the range is between [-83, 0], it is divided into NS; when e The range is between [0, 167]. ec When the range is between [0, 83], it is classified as PS; when e The range is between [167, 333]. ec When the range is between [83, 167], it is classified as PM; when e The range is between [333, 500]. ec When the range is between [167, 250], it is classified as PB; By introducing a scaling factor and performing fuzzing processing. and The domain range is quantized to [0.5, 2.0], and the fuzzy level is represented by a trapezoidal membership function, with the core interval being a flat-topped segment that gradually decreases towards zero on both sides. The membership function is defined as follows: NB (Negative Large): Coverage range is [0.5, 0.8], core interval is [0.6, 0.7], decreasing on both sides.

[0026] NM (Negative Midpoint): Coverage range is [0.7, 1.0], core interval is [0.8, 0.9], decreasing on both sides.

[0027] NS (Negative Small): Coverage range is [0.9, 1.2], core interval is [1.0, 1.1], decreasing on both sides.

[0028] PS (smallest): Coverage range is [1.1, 1.4], core interval is [1.2, 1.3], decreasing on both sides.

[0029] PM (center): Coverage range is [1.3, 1.6], core range is [1.4, 1.5], decreasing to both sides.

[0030] PB (Zhengda): Coverage range is [1.6, 2.0], core interval is [1.7, 1.9], decreasing on both sides.

[0031] Each fuzziness level The membership degree is calculated as follows: NB: (9) NM: (10) NS: (11) PS: (12) PM: (13) PB: (14) The domain range is quantized to [0.5, 1.0]. The fuzzy level is represented by a triangular membership function, with the center point corresponding to a membership degree of 1, and the degree gradually decreasing towards zero on both sides. The membership function is defined as follows: NB (Negative Large): Center point 0.55, coverage area [0.5, 0.6], decreasing on both sides.

[0032] NM (Negative Center): Center point 0.65, coverage area [0.58, 0.72], decreasing on both sides.

[0033] NS (Negative Small): Center point 0.75, coverage range [0.7, 0.8], decreasing on both sides.

[0034] PS (smallest): Center point 0.85, coverage area [0.78, 0.92], decreasing on both sides.

[0035] PM (center): Center point 0.95, coverage area [0.9, 1.0], decreasing to both sides.

[0036] PB (Zhengda): Center point 0.99, coverage area [0.98, 1.0], decreasing on both sides.

[0037] Each fuzziness level The membership degree is calculated as follows: NB: (15) NM: (16) NS: (17) PS: (18) PM: (19) PB: (20) Output variables and The linguistic values ​​are described using B (limiting expansion of the universe of discourse), MB (strong expansion of the universe of discourse), M (no expansion of the universe of discourse), MS (weak expansion of the universe of discourse), and S (minimal expansion of the universe of discourse); output variables The linguistic variables are described using Z (uncompressed universe of discourse), MS (moderately compressed universe of discourse), and S (limitedly compressed universe of discourse).

[0038] Fuzzy rules are typically expressed in the form of "if...then...". The design rules for the universe of discourse scaling factor are as follows: If both the error in the number of sparks and the rate of change of the error are positive, the output voltage rise speed needs to be significantly increased to improve the control effect of the change in the number of sparks. If both the error in the number of sparks and the rate of change of the error are negative, the rise rate of the output voltage should be reduced to control the decrease in the number of discharge sparks. If the error in the number of sparks is positive but the rate of change of the error is negative, the rise rate of the output voltage should be appropriately reduced to make the trend of the number of discharge sparks more similar. If the error in the number of sparks is negative but the rate of change of error is positive, it is necessary to significantly reduce the rise rate of the output voltage to control the increase in the number of discharge sparks.

[0039] and The specific rules are as follows: If ( e is NB) and ( ec is NB) then ( is B); If ( e is NB) and ( ec is NM)then( is B); If ( e is NB) and ( ec is NS) then ( is B); If ( e is NB) and ( ec is PS) then ( is MS); If ( e is NB) and ( ec is PM) then ( is S); If ( e is NB) and ( ec is PB)then( is S); If ( eis NM)and( ec is NB)then( is B); If( e is NM)and( ec is NM)then( is MB); If( e is NM)and( ec is NS)then( is M); If( e is NM)and( ec is PS)then( is S); If( e is NM)and( ec is PM)then( is S); If( e is NM)and( ec is PB)then( is S); If( e is NS)and( ec is NB)then( is MB); If( e is NS)and( ec is NM)then( is M); If( e is NS)and( ec is NS)then( is MS); If( e is NS)and( is PS)then( is S); If( e is NS)and( ec is PM)then( is S); If( e is NS)and( ec is PB)then( is S); If( e is PS)and(ec is NB)then( is S); If( e is PS)and( ec is NM)then( is S); If( e is PS)and( ec is NS)then( is S); If( e is PS)and( ec is PS)then( is M); If( e is PS)and( ec is PM)then( is MB); If( e is PS)and( ec is PB)then( is B); If( e is PM)and( ec is NB)then( is S); If( e is PM)and( ec is NM)then( is S); If( e is PM)and( ec is NS)then( is S); If( e is PM)and( is PS)then( is M); If( e is PM)and( ec is PM)then( is MB); If( e is PM)and( ec is PB)then( is B); If( e is PB)and( ecis NB)then( is S); If( e is PB)and( ec is NM)then( is S); If( e is PB)and( ec is NS)then( is MS); If( e is PB)and( ec is PS)then( is MB); If( e is PB)and( ec is PM)then( is B); If( e is PB)and( ec is PB)then( is B); The specific rules are as follows: If( e is NB)and( ec is NB)then( is Z); If( e is NB)and( ec is NM)then( is Z); If( e is NB)and( ec is NS)then( is MS); If( e is NB)and( ec is PS)then( is S); If( e is NB)and( ec is PM)then( is S); If( e is NB)and( ec is PB)then( is S); If( eis NM)and( ec is NB)then( is Z); If( e is NM)and( ec is NM)then( is MS); If( e is NM)and( ec is NS)then( is MS); If( e is NM)and( ec is PS)then( is S); If( e is NM)and( ec is PM)then( is S); If( e is NM)and( ec is PB)then( is S); If( e is NS)and( ec is NB)then( is MS); If( e is NS)and( ec is NM)then( is MS); If( e is NS)and( ec is NS)then( is S); If( e is NS)and( ec is PS)then( is S); If( e is NS)and( ec is PM)then( is S); If( e is NS)and( ec is PB)then( is S); If( e is PS)and(ec is NB)then( is S); If( e is PS)and( ec is NM)then( is S); If( e is PS)and( ec is NS)then( is S); If( e is PS)and( ec is PS)then( is MS); If( e is PS)and( ec is PM)then( is MS); If( e is PS)and( ec is PB)then( is Z); If( e is PM)and( ec is NB)then( is S); If( e is PM)and( ec is NM)then( is S); If( e is PM)and( ec is NS)then( is S); If( e is PM)and( ec is PS)then( is MS); If( e is PM)and( ec is PM)then( is Z); If( e is PM)and( ec is PB)then( is Z); If( e is PB)and( ecis NB) then ( is S); If ( e is PB) and ( ec is NM)then( is S); If ( e is PB) and ( ec is NS) then ( is MS); If ( e is PB) and ( ec is PS) then ( is Z); If ( e is PB) and ( ec is PM) then ( is Z); If ( e is PB) and ( ec is PB)then( is Z); Based on the error in the number of discharge sparks from the input. e and its error change rate ec Scaling factor of the universe of discourse , Design new fuzzy rules, as shown in Table 1. The fuzzy rules are shown in Table 2, forming a fuzzy controller with two inputs and three outputs.

[0040] Table 1 , Fuzzy rule table

[0041] Table 2 Fuzzy rule table

[0042] Example 5 The specific process of step 6 is as follows: Step 6.1, based on the fuzzy rule base and the fuzzified input signal, fuzzy logic operations are used to derive the fuzzy control output. The Mamdani inference method is adopted, and its basic inference process is: for each fuzzy rule, the applicability of the rule is calculated based on the membership degree of the input signal and the preconditions of the rule; then, the applicability of all rules is combined with the corresponding fuzzy conclusion set to obtain a fuzzy control output set.

[0043] Step 6.1.1, determine the current... e andec Determine whether the fuzzy membership degree satisfies the rule preconditions and calculate the applicability of the activation rule.

[0044] Rule prerequisites: If a certain rule's " e The corresponding fuzzy set e Membership degree > 0, and " ec The corresponding fuzzy set ec ec ec If the membership degree is greater than 0, the rule is activated and incorporated into subsequent reasoning; If the membership degree of any input variable to the fuzzy set of the rule premises is 0, then the rule does not match and is excluded from the reasoning process.

[0045] For each activation rule, the applicability (rule strength) is calculated using the minimum (min) operation, as shown in the formula: (twenty one) This represents the applicability of a single activation rule, with a value range of [0,1]. , It represents the membership degree of a fuzzy subset of the variable universe of discourse.

[0046] Step 6.1.2, for each activation rule, the three output variables ( , , ), respectively using their applicability The upper bound of the fuzzy set of conclusions is truncated using the following formula: (twenty two) in: The original membership function of the fuzzy set of the output variable conclusion ( , For trapezoidal membership functions, (Membership function of triangle) This is the truncated membership function, retaining only the membership values ​​≤ applicability from the original membership function. The part that exceeds the limit is removed to determine the degree of matching rule adaptation.

[0047] The maximum (max) value is taken from the fuzzy sets of conclusions after truncating all activation rules, and then a unified fuzzy control output set for the output variable is synthesized using the following formula: (twenty three) in, This represents the set of synthesized fuzzy control outputs.

[0048] The fuzzy reasoning process, through the fuzziness and uncertainty of the input signal, can simulate the fuzzy reasoning method of humans and ultimately obtain the fuzzy control output.

[0049] Step 6.2: Defuzzify the fuzzy control output using the centroid method. The defuzzification formula is as follows: (twenty four) in, express The first in i The fuzzy control output values ​​corresponding to each fuzzy subset. This represents the scaling factor of the output universe of discourse. This represents the output value after defuzzification.

[0050] By defuzzifying, the fuzzy control output is converted into an output reference voltage signal. u .

[0051] Example 6 The feedback spark quantity characteristic diagram under VUFC is shown in Figure 4(a). It can be seen from the figure that the spark curve output under VUFC control can fit the expected spark curve well and the error is small. The output voltage of VUFC control is shown in Figure 4(b). VUFC control can adjust the output voltage according to the error between the current expected spark quantity and the sampled spark quantity. The output voltage rises steadily and does not fluctuate significantly.

[0052] This invention introduces a scaling factor to adjust the domain of input and output variables in real time according to the system state; furthermore, it obtains the output reference voltage according to fuzzy rules, so that the number of discharge sparks can maintain high control accuracy at different reaction stages, thus ensuring the stability and consistency of the ceramic layer process.

Claims

1. A method for controlling the number of sparks in liquid-phase discharge plasma, characterized in that: Specifically, the steps include the following: Step 1: Acquire an image of the discharge spark and obtain the number of sparks in the image; Step 2: Calculate the deviation and rate of change of the number of sparks obtained from the set number of sparks during the discharge process. Step 3: Based on the spark quantity deviation and rate of change, a dynamic scaling factor is introduced to adjust the domain of discourse; Step 4: Fuzzyize the scaling factors of the input and output variables and establish membership functions; Step 5: Design fuzzy rules for the dynamic scaling factor; Step 6: Perform fuzzy inference based on the membership degree and fuzzy rules of the scaling factor signal, and then perform defuzzification; Step 7: Adjust the number of discharge sparks based on the output reference voltage obtained from the defuzzification.

2. The method for controlling the number of liquid-phase discharge plasma sparks according to claim 1, characterized in that: The specific process of step 2 is as follows: Use the set target number of sparks Subtract the number of discharge sparks during the liquid phase discharge process ,get t Spark count deviation at any moment e ( t and the corresponding rate of change of deviation ec ( t ), （1） （2） in, e ( t () represents the deviation in the number of sparks at time t. e ( t -1) represents the spark count deviation at time t-1.

3. The method for controlling the number of liquid-phase discharge plasma sparks according to claim 2, characterized in that: The specific process of step 3 is as follows: the scaling factor of the input variable for controlling the number of constant discharge sparks is... The scaling factor of the output variable is The basic universe of discourse of the scaled input variables X ( e )and X ( ec and the fundamental universe of discourse of the output variable Y(u) They can be represented as follows: （3） in, and This is the scaling factor for the input variables.

4. The method for controlling the number of liquid-phase discharge plasma sparks according to claim 3, characterized in that: The specific process of step 4 is as follows: Step 4.1, set the quantization domain of the scaling factors of the two input variables. X =[ x l , x h The quantization domain of the output variable scaling factor. Y =[ y l , y h Error in the number of discharge sparks e Quantification factor K e Error change rate ec Quantification factor K ec and output reference voltage u Quantification factor K u The calculation formula is: （4） （5） （6） in, and The quantization domain is [0.5, 2.0], and the corresponding fuzzy set is {S, MS, M, MB, B}; The quantization domain is [0.5, 1.0], and the corresponding fuzzy set is {S, MS, Z}; Step 4.2, and The trapezoidal membership function consists of four parameters. a , b , c , d Define, where a < b < c < d ,but and Membership function Represented as: （7） The triangular membership function consists of three parameters. a , b , c Defined, then Membership function Represented as: （8） in, a < b < c .

5. The method for controlling the number of liquid-phase discharge plasma sparks according to claim 4, characterized in that: The specific process of step 5 is as follows: design the input and output variables of the fuzzy controller: Input variables e and ec The six linguistic variables {NB, NM, NS, PS, PM, PB} are used to describe the values, which correspond to negative large, negative medium, negative small, positive small, positive medium, and positive large, respectively. e The range is between [-500, 500]. ec The range is between [-250, 250], and the range of language variables is divided as follows: when e The range is between [-500, -333]. ec When the range is between [-250, -167], it is classified as NB; when e The range is between [-333, -167]. ec When the range is between [-167, -83], it is classified as NM; when e The range is between [-167, 0]. ec When the range is between [-83, 0], it is divided into NS; when e The range is between [0, 167]. ec When the range is between [0, 83], it is classified as PS; when e The range is between [167, 333]. ec When the range is between [83, 167], it is classified as PM; when e The range is between [333, 500]. ec When the range is between [167, 250], it is classified as PB; By introducing a scaling factor and performing fuzzing processing. and The domain range is quantized to [0.5, 2.0], and the membership function is defined as follows: NB: Coverage range is [0.5, 0.8], core range is [0.6, 0.7], decreasing to both sides; NM: Coverage range is [0.7, 1.0], core interval is [0.8, 0.9], decreasing on both sides; NS: Coverage range is [0.9, 1.2], core interval is [1.0, 1.1], decreasing on both sides; PS: The coverage area is [1.1, 1.4], the core interval is [1.2, 1.3], and the range decreases on both sides; PM: Coverage range is [1.3, 1.6], core range is [1.4, 1.5], decreasing on both sides; PB: Coverage range is [1.6, 2.0], core interval is [1.7, 1.9], decreasing on both sides; Each fuzziness level The membership degree is calculated as follows: NB: (9) NM: (10) NS: (11) PS: (12) PM: (13) PB: (14) The domain range is quantized to [0.5, 1.0], and the membership function is defined as follows: NB: Center point 0.55, coverage area [0.5, 0.6], decreasing on both sides; NM: Center point 0.65, coverage area [0.58, 0.72], decreasing on both sides; NS: Center point 0.75, coverage area [0.7, 0.8], decreasing on both sides; PS: Center point 0.85, coverage area [0.78, 0.92], decreasing on both sides; PM: Center point 0.95, coverage area [0.9, 1.0], decreasing on both sides; PB: Center point 0.99, coverage area [0.98, 1.0], decreasing on both sides; Each fuzziness level The membership degree is calculated as follows: NB: (15) NM: (16) NS: (17) PS: (18) PM: (19) PB: (20) Among them, output variables and The language values ​​are described using B, MB, M, MS, and S; output variables The linguistic variables are described by Z, MS, and S.

6. The method for controlling the number of sparks in liquid-phase discharge plasma according to claim 5, characterized in that: The specific process of step 6 is as follows: Step 6.1: Based on the fuzzy rule base and the fuzzified input signal, fuzzy logic operations are used to derive the fuzzy control output; Step 6.2: Use the centroid method to defuzzify the fuzzy control output.

7. The method for controlling the number of sparks in liquid-phase discharge plasma according to claim 6, characterized in that: The specific process of step 6.1 is as follows: Step 6.1.1, determine the current... e and ec The fuzzy membership degree is evaluated to determine whether it meets the rule preconditions and to calculate the applicability of the activation rule. Specifically: Rule prerequisites: If a certain rule's " e The corresponding fuzzy set e Membership degree > 0, and " ec The corresponding fuzzy set ec If the membership degree is greater than 0, the rule is activated and incorporated into subsequent reasoning; If the membership degree of any input variable to the fuzzy set of the rule premises is equal to 0, then the rule does not match and is excluded from the reasoning process; For each activation rule, the applicability is calculated using the minimum value operation, with the following formula: (21) in, This represents the applicability of a single activation rule, with a value range of [0,1]. , Represents the membership degree of a fuzzy subset with a variable universe of discourse; Step 6.1.2, for the three output variables of each activation rule , , Using applicability respectively The upper bound of the fuzzy set of conclusions is truncated using the following formula: (22) in, The original membership function of the fuzzy set of the output variable conclusion. , For trapezoidal membership functions, Let the membership function of the triangle be . This is the truncated membership function, retaining only the membership values ​​≤ applicability from the original membership function. The portion exceeding the limit is removed to determine the matching rule's fit. The maximum value of the fuzzy sets of conclusions after truncating all activation rules is used to synthesize a unified fuzzy control output set for the output variable, as shown in the formula: (23) in, This represents the set of synthesized fuzzy control outputs.

8. The method for controlling the number of sparks in liquid-phase discharge plasma according to claim 7, characterized in that: In step 6.2, the formula for defuzzification is as follows: (24) in, express The first in i The fuzzy control output values ​​corresponding to each fuzzy subset. This represents the scaling factor of the output universe of discourse. This represents the output value after defuzzification.

Images