Improved grey GM (1, 1) prediction method and system for output current of EAST fast control power supply

Abstract

The invention relates to the technical field of EAST fast control power supply operation, and provides an EAST fast control power supply output current improved gray GM (1, 1) prediction method comprising the following steps: sampling to obtain an original sequence X (0) and generate an accumulation sequence X (1) (k), segmenting the accumulation sequence in a gray GM (1, 1) prediction model background value construction to obtain a background value, and calculating the gray GM (1, 1) prediction model; when the concavity and convexity of the predicted value sequence curve and the concavity and convexity of the output current curve are different, reconstructing an accumulation sequence by using a transformation operator, and recalculating a background value, a grey development coefficient and a grey amount; replacing the old value in the original sequence with the predicted output current of the next switching period to update the original sequence, further predicting the output current of the next two switching periods, and compensating the inherent time delay of digital lag control; accurate prediction of the output current of the EAST fast control power supply is realized, the response speed of the output current is accelerated, and the balance control performance of plasma vertical displacement is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of EAST fast control power supply operation, and in particular to an EAST fast control power supply output current improved grey GM (1,1) prediction control method and a system thereof. Background Art

[0002] The Experimental Advanced Superconducting Tokamak (EAST) is an important engineering device for realizing controlled nuclear fusion. The main function of the fast-controlled power supply is to excite the load coil according to the reference signal output current to ensure that the plasma can quickly achieve magnetic confinement balance control in the vertical displacement, avoiding the plasma hitting the vacuum wall and causing discharge failure.

[0003] The plasma vertical displacement active feedback control system detects the vertical displacement of the plasma in real time, and calculates the reference signal required to maintain the vertical displacement balance of the plasma in real time. The EAST fast-control power supply tracks the given reference signal and quickly outputs current to excite the load coil to achieve fast closed-loop feedback control of the plasma in the vertical direction, ensuring the rapidity of the vertical displacement of the plasma. The reference signal is linearly corresponding to the output current of the fast-control power supply. The given reference signal of ±10V corresponds to the output current of the EAST fast-control power supply branch of ±1500A. When the plasma is offset in vertical displacement, in order to quickly pull the plasma back to the vertical equilibrium position to avoid plasma collision and fragmentation and discharge failure, the EAST fast-control power supply is required to have a faster dynamic response speed of the output current, which places higher requirements on the dynamic response performance of the power supply.

[0004] The EAST fast-control power supply adopts a digital hysteresis control method. The speed of the output current is mainly limited by the voltage applied to the load coil at the output end and the delay time of the system itself. The higher the voltage level applied to the load coil at the output end and the shorter the delay time of the digital hysteresis control system, the faster the output current is established. In actual engineering applications, in order to further improve the output current establishment speed of the EAST fast-control power supply, it is necessary to further improve the voltage level at the output end of the load coil. This also requires that the voltage and current resistance capabilities of the power devices in the EAST fast-control power supply need to be further improved. However, in order to avoid a significant increase in switching losses, the switching frequency of the power switching devices needs to be further reduced. The lower switching frequency brings a larger system control delay, resulting in a slower output current establishment speed of the EAST fast-control power supply. The need to increase the output voltage level and the constraint of reducing the switching frequency bring contradictory constraints to the rapid establishment of the output current of the EAST fast-control power supply, making it difficult to further improve the dynamic response speed of the output current.

[0005] Predictive control that predicts the output current trajectory in advance is an effective means to improve the output current response speed of the EAST fast-control power supply. The gray GM (1,1) prediction only requires a small amount of known target information to achieve the prediction of the target quantity. For example, the Chinese invention patent "EAST fast-control power supply output current prediction method based on improved gray GM (1,1) model prediction" with publication number CN111523700A discloses an improved gray GM (1,1) model prediction method. According to the monotonicity of the output current of the EAST fast-control power supply, the original sequence of the gray prediction is divided into two phase sequences, namely, the rising phase sequence and the falling phase sequence. A non-equally spaced gray GM (1,1) prediction model is established in the rising phase sequence, and an equidistant gray GM (1,1) prediction model is established in the falling phase sequence, so as to achieve accurate and rapid control of the output current of the EAST fast-control power supply. The Chinese invention patent "Transformer fault prediction method based on gray GM (1,1) model" with publication number CN113269350A optimizes the errors existing in the GM (1,1) model in turn, and obtains a total of eight optimized prediction schemes according to the combination arrangement, which meets the diversified needs. The Chinese invention patent with publication number CN113592131A, "A method for predicting the annual generation of construction waste based on an improved grey GM (1,1) model", divides the original sequence into different stage sequences, establishes a non-uniformly spaced improved grey GM (1,1) model, and realizes precise positioning and rapid control of construction waste monitoring and management.

[0006] The above patent application modified the grey GM(1,1) prediction model and established a non-uniformly spaced improved grey GM(1,1) model or an optimized prediction scheme. However, in the output current prediction and control of the EAST fast-control power supply, there are still problems such as the inconsistency between the convexity and concavity of the grey GM(1,1) prediction value sequence and the actual output current curve, the background value construction process is too complicated and even has large errors, the old information of the original sequence seriously affects the prediction accuracy, and the digital hysteresis control delay seriously affects the output current response speed. These problems result in a decrease in the prediction accuracy of the EAST fast-control power supply output current and seriously affect the output current control performance. Summary of the invention

[0007] The technical problem to be solved by the present invention is how to reduce the prediction error of the output current of the EAST fast-control power supply, thereby ensuring the rapidity and accuracy of the output current of the EAST fast-control power supply.

[0008] The present invention solves the above technical problems through the following technical means:

[0009] The present invention provides an improved grey GM (1,1) prediction method for output current of an EAST fast-controlled power supply, comprising the following steps:

[0010] S1. Sample the output current of the EAST fast control power supply to obtain the original sequence X of the grey GM (1,1) prediction model composed of the output current sampling data in the last four switching cycles. (0) , the expression is:

[0011] X (0) =[X (0) (1),X (0) (2),...,X (0) (n)] (1)

[0012] S2, based on the original sequence X (0) , generate the cumulative sequence X (1) (k), where k represents the switching period, and then combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k);

[0013] S3, based on the background value sequence Z obtained in step S2 (1) (k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b; judge whether the background value sequence Z needs to be reconstructed according to the size of the gray development coefficient a (1) (k) and recalculate the grey development coefficient a and the grey quantity b;

[0014] S4, based on the gray development coefficient a, gray quantity b and original sequence X obtained in step S3 (0) , predict the output current i(k+1) of the next switching cycle, and replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) ;

[0015] S5, based on the updated original sequence X (0) , repeat steps S2 to S4 to predict the output current i(k+2) of the next switching cycle again;

[0016] S6. Take i(k+2) as the latest predicted value of the output current; subtract the signal corresponding to i(k+2) from the reference signal to obtain the output current error; when the output current error is small, adjust the proportional control parameter of the EAST fast-control power supply to be smaller; when the output current error is large, adjust the proportional control parameter of the EAST fast-control power supply to be larger to adjust the duty cycle of the power switch tube.

[0017] Furthermore, the cumulative sequence expression described in step S2 is:

[0018]

[0019] Furthermore, the cumulative sequence X is combined with the calculus idea as described in step S2. (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k) Specifically:

[0020] The accumulated sequence X (1) (k) The curve is divided into l rectangles, and the area of ​​l rectangles is used to approximate the integral value of the curve. (1) (k) After segmentation, the background value sequence Z is approximately calculated (1) (k), the background value sequence expression is obtained as:

[0021]

[0022] Where k represents the switching period.

[0023] Furthermore, the gray development coefficient a and gray amount b are calculated in step S3, and the specific calculation expressions are:

[0024] X n =[X (0) (2)X (0) (3)X (0) (4)] T (5)

[0025]

[0026] Furthermore, in step S3, the gray development coefficient a is used to determine whether the background value sequence Z needs to be reconstructed. (1) (k) and update the calculation of the gray development coefficient a and the gray amount b, specifically:

[0027] When a>0, the background value sequence Z is retained (1) (k), and retain the calculated gray development coefficient a and gray amount b;

[0028] When a<0, the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), and based on the reconstructed cumulative sequence X Δ(1) (k) Recalculate the background value sequence Z (1) (k), and then based on the recalculated background value sequence Z (1) (k) Update and calculate the gray development coefficient a and gray amount b.

[0029] Furthermore, the cumulative sequence X reconstructed by the transformation operator Δ(1) (k) specifically:

[0030] Reconstruct the cumulative sequence X Δ(1) In the process (k), the original sequence error caused by the inconsistency of concavity and convexity is first calculated as

[0031] Δ(k+1)=X (1) (nk)-X (1) (nk-1),k=0,1,...,n-2 (8)

[0032] The reconstructed cumulative sequence is

[0033]

[0034] Furthermore, the output current i(k+1) of the next switching cycle described in step S4 is predicted by:

[0035]

[0036] Furthermore, the updated original sequence X in step S4 is (0) The expression is

[0037] X (0) =[X (0) (2),...,X (0) (n),i(k+1)]. (11)

[0038] The present invention also provides an EAST fast-controlled power supply output current improved grey GM (1,1) prediction system, wherein the system adopts the above method when running, and comprises the following modules:

[0039] The sampling module is used to sample the output current of the EAST fast control power supply and obtain the original sequence X of the grey GM (1,1) prediction model composed of the output current sampling data in the last four switching cycles. (0) , the expression is:

[0040] X (0) =[X (0) (1),X (0) (2),...,X (0) (n)] (1)

[0041] Background value sequence calculation module is used to calculate the background value sequence based on the original sequence X (0) , generate the cumulative sequence X (1) (k), where k represents the switching period, and then combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k);

[0042] Gray parameter calculation module, used for background value sequence Z obtained based on background value sequence calculation module (1)(k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b; judge whether the background value sequence Z needs to be reconstructed according to the size of the gray development coefficient a (1) (k) and recalculate the grey development coefficient a and the grey quantity b;

[0043] The first current prediction module is used to predict the gray development coefficient a, gray quantity b and the original sequence X obtained in step S3. (0) , predict the output current i(k+1) of the next switching cycle, and replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) ;

[0044] The second current prediction module is used to predict the current based on the updated original sequence X (0) , repeatedly run the background value sequence calculation module, the gray parameter calculation module and the first current prediction module to predict the output current i(k+2) of the next switching cycle again;

[0045] The control module is used to use i(k+2) as the latest predicted value of the output current; the signal corresponding to i(k+2) is subtracted from the reference signal to obtain the output current error; when the output current error is small, the proportional control parameter of the EAST fast control power supply is adjusted to be smaller; when the output current error is large, the proportional control parameter of the EAST fast control power supply is adjusted to be larger to adjust the duty cycle of the power switch tube.

[0046] Furthermore, the grey parameter calculation module includes a grey development coefficient a determination unit, and its specific operation mode is as follows:

[0047] When a>0, the background value sequence Z is retained (1) (k), and retain the calculated gray development coefficient a and gray amount b;

[0048] When a<0, the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), and based on the reconstructed cumulative sequence X Δ(1) (k) Recalculate the background value sequence Z (1) (k), and then based on the recalculated background value sequence Z (1) (k) Update and calculate the gray development coefficient a and gray amount b.

[0049] The advantages of the present invention are:

[0050] (1) This invention improves the background value sequence Z obtained by the median theorem in traditional grey prediction. (1) (k), combined with the idea of ​​calculus, the cumulative sequence X (1)(k) The curve is divided into l rectangles to replace the right-angled trapezoid in the mean value theorem. The area of ​​l rectangles is used to approximate the integral value of the curve. The cumulative sequence X (1) (k) After segmentation, the background value sequence Z is approximately calculated (1) (k) is used to reduce the error caused by the median theorem, so that the background value sequence is more consistent with the law of the grey GM (1,1) prediction model and the prediction accuracy of the grey GM (1,1) prediction model is improved.

[0051] (2) The present invention is based on the background value sequence Z (1) (k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b, and further determine the value of a; when a>0, the concavity of the predicted value sequence curve obtained by prediction is the same as the concavity of the actual output current curve, and the background value sequence Z is retained. (1) (k), and retain the calculated gray development coefficient a and gray quantity b; when a<0, the concavity of the predicted value sequence curve obtained by prediction is different from the concavity of the actual output current curve, then the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), then based on the reconstructed cumulative sequence X Δ(1) (k) and calculus ideas to recalculate the background value sequence Z (1) (k), and according to the updated background value sequence Z (1) (k) Recalculate the gray development coefficient a and the gray quantity b; determine whether the values ​​of a and b need to be recalculated based on the values ​​of the gray development coefficient a and the gray quantity b, thereby ensuring that the predicted value sequence curve is consistent with the characteristic curve of the actual output current, thereby further improving the prediction accuracy of the output current.

[0052] (3) The present invention promptly discards the old original sequence value and updates the original sequence X (0) , the gray development coefficient a and gray amount b are calculated again, the influence of old information is eliminated, and a rolling two-step output current prediction is realized to ensure the output current prediction accuracy, especially when the reference signal mutates, the output current prediction accuracy is further improved.

[0053] (4) In step six of the present invention, the current i(k+2) is used as the latest predicted value of the output current, which compensates for the inherent delay of the EAST fast-controlled power supply digital hysteresis control system, so that the current value sampled in the k+1th cycle is actually the current output value obtained by the system under the control of the control quantity u(k+1), which is consistent with the value expected by the system, thereby improving the dynamic response speed of the output current and the tracking accuracy of the output current.

[0054] (5) Combining the gray GM (1,1) prediction model to make advance predictions on the output current trajectory, it is possible to achieve faster and more accurate output of the output current without increasing the power level of the power supply, significantly improving the balance control performance of the plasma displacement in the vertical direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a schematic diagram of the prediction process of the grey GM (1,1) model of the overall output current in the example of the present invention;

[0056] Figure 2 A schematic diagram of a background value sequence for segmentation and approximate calculation in combination with the idea of ​​calculus in an example of the present invention;

[0057] Figure 3 A schematic diagram of compensating for digital hysteresis control delay in an example of the present invention;

[0058] Figure 4 This is a schematic diagram of the predicted output current tracking given signal waveform of the EAST fast control power supply gray GM (1,1) in the example of the present invention. DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] Example 1

[0061] This example provides an improved grey GM(1,1) prediction method for the output current of an EAST fast-controlled power supply. Figure 1 As shown, the following steps are included:

[0062] S1. The output current of the EAST fast control power supply is sampled through a 1:5000 current sensor and an ADC digital sampling chip to obtain the original sequence X of the grey GM (1,1) prediction model consisting of the output current sampling data in the last four switching cycles. (0) , the original sequence expression is

[0063] X (0) =[X (0) (1),X (0) (2),...,X (0) (n)] (1)

[0064] Among them, n can be 4, X (0) (1)~X(0) (n) represents the first 4 sampling data points, and X will be deleted first when the original sequence is updated later. (0) (1), the remaining data will be moved forward, and the newly added data will be placed at the end of the data group to ensure the time validity of the original sequence.

[0065] S2, based on the original sequence X (0) , generate the cumulative sequence X (1) (k), the generated cumulative sequence expression is

[0066]

[0067] Among them, k represents the switching cycle, and combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k).

[0068] In traditional grey prediction, the median theorem is used as a simple approximate substitute calculation, but there is a large error. The background value constructed is

[0069]

[0070] In this embodiment, the cumulative sequence X is combined with the idea of ​​calculus (1) (k) The curve is divided into l rectangles to replace the right-angled trapezoid in the mean value theorem, such as Figure 2 As shown, the area of ​​l rectangles is used to approximate the integral value of the curve, and the cumulative sequence X (1) (k) After segmentation, the background value sequence Z is approximately calculated (1) (k), in order to reduce the error caused by the median theorem, the background value sequence expression obtained by segmentation approximate calculation is:

[0071]

[0072] S3, based on background value sequence Z (1) (k) and the original current sampling data, calculate the gray development coefficient a and gray amount b, and the expression is:

[0073] X n =[X (0) (2)X (0) (3)X (0) (4)] T (5)

[0074]

[0075] In traditional grey prediction, the output can be predicted by calculating the development coefficient a and the grey quantity b. However, in this embodiment, in order to ensure that the predicted value sequence curve obtained by prediction is consistent with the characteristics of the actual output current curve and to ensure the accuracy of the prediction, it is necessary to further determine the value of a; when a>0, the trend of the predicted value sequence curve obtained by prediction will show the characteristics of an increasing convex function and a decreasing concave function, which is the same as the characteristic curve of the actual output current, and the background value sequence Z obtained is retained. (1) (k), and retain the calculated gray development coefficient a and gray quantity b; when a<0, the trend of the predicted value sequence curve obtained by prediction will show the characteristics of increasing concave function and decreasing convex function, which is different from the characteristic curve of the actual output current. Then, the cumulative sequence X is reconstructed through the transformation operator. Δ(1) (k); Reconstruct the cumulative sequence X Δ(1) In the process (k), the original sequence error caused by the inconsistency of concavity and convexity is first calculated as

[0076] Δ(k+1)=X (1) (nk)-X (1) (nk-1),k=0,1,…,n-2 (8)

[0077] The reconstructed cumulative sequence is

[0078]

[0079] Based on the reconstruction of the cumulative sequence X Δ(1) (k) and calculus ideas to update the background value sequence Z (1) (k), according to the updated background value sequence Z (1) (k) Recalculate the gray development coefficient a and gray amount b.

[0080] S4, according to the gray development coefficient a and gray amount b and the original sequence X (0) , predict the output current i(k+1) of the next switching cycle, and its prediction expression is

[0081]

[0082] And replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) , the newly added data will be placed at the end of the data group to ensure the time validity of the original sequence. The updated original sequence X (0) The expression is

[0083] X (0) =[X (0) (2),...,X (0) (n),i(k+1)] (11)

[0084] S5, based on the updated original sequence X (0) , repeat steps S2 to S4, and predict the output current i(k+2) of the next switching cycle again; when the reference signal suddenly changes at a certain moment, the original sequence X (0) The output current value before the mutation exists in , which reduces the prediction accuracy of the output current. In order to improve the prediction accuracy of the output current during the reference signal mutation process, the old original sequence value should be discarded in time to realize rolling grey prediction. Therefore, according to formula (10), the output current i(k+2) of the next switching cycle is predicted, and two predicted currents i(k+1) and i(k+2) after the kth cycle are obtained, thereby eliminating the influence of the old information, realizing the rolling two-step output current prediction, and ensuring the output current prediction accuracy.

[0085] S6. Use the current i(k+2) as the latest predicted value of the output current to compensate for the inherent delay of the digital hysteresis control, and subtract the signal corresponding to i(k+2) from the reference signal to obtain the output current error; when the output current error is small, adjust the proportional control parameter of the EAST fast-control power supply to be smaller; when the output current error is large, adjust the proportional control parameter of the EAST fast-control power supply to be larger, so as to obtain the optimal duty cycle of the power switch tube and realize fast and accurate prediction and control of the output current of the EAST fast-control power supply.

[0086] like Figure 3 As shown, T sis the switching cycle of the switch tube. The EAST fast-control power supply digital hysteresis control system has a control delay of at least one switching cycle. According to the gray prediction model, the output current i(k+1) at time k+1 is predicted in the kth cycle, and the control quantity u(k) required by the EAST fast-control power supply switch tube is calculated. The control module loads the PWM wave under the control quantity u(k-1) calculated in the previous cycle to drive the switch tube. Therefore, the current value sampled in the k+1th cycle is actually the current output value obtained by the system under the control of the control quantity u(k-1), which deviates from the value expected by the system. In order to compensate for the control delay of the EAST fast-control power supply digital hysteresis control system and improve the output current prediction accuracy during the reference signal mutation process, the old The original sequence value realizes rolling grey prediction, so in the kth cycle, a rolling two-step output current prediction is realized to obtain i(k+2) and calculate the control quantity u(k+2), and u(k+2) is loaded in the k+1th cycle. Then, the current value sampled in the k+2th cycle is actually the current output value obtained by the system under the control of the control quantity u(k+2), which is consistent with the value expected by the system; the signal corresponding to i(k+2) is subtracted from the reference signal to obtain the output current error to ensure that the predicted current quickly and accurately tracks the reference signal, and the proportional control parameter size of the EAST fast-control power supply is adjusted to obtain the optimal duty cycle of the power switch tube to ensure that the predicted current accurately tracks the reference signal, thereby realizing fast and accurate prediction and control of the output current of the EAST fast-control power supply.

[0087] In this embodiment, the improved grey GM (1,1) prediction method for the output current of the EAST fast control power supply is simulated and analyzed, and the simulation parameters are: the DC supply voltage of the EAST fast control power supply is 540V, the load inductance value is 160μH, the load inductance internal resistance value is 0.08Ω, the switching frequency of the switching device is 5kHz, the reference signal is a 100Hz square wave of ±10V, and the output current rating is ±1500A. The improved grey GM (1,1) prediction control method is used to realize the prediction control of the output current of the EAST fast control power supply. Figure 4 As shown, analysis of the simulation waveform shows that when the reference signal is a 100Hz AC square wave of ±10V, the EAST fast-control power supply can quickly output a 100Hz AC square wave current of ±1500A, and the proportional relationship between the reference signal value and the output current value of the EAST fast-control power supply fully meets the requirements for the normal operation of the EAST fast-control power supply. The simulation waveform shows that the present invention can enable the EAST fast-control power supply to quickly and accurately follow the reference signal to output the corresponding current, and can achieve good plasma vertical displacement balance control performance.

[0088] Example 2

[0089] It should be further explained that, based on the same inventive concept, this embodiment provides an EAST fast-control power supply output current improved grey GM (1,1) prediction system, which executes the method described in embodiment 1 during operation, including the following modules:

[0090] The sampling module is used to sample the output current of the EAST fast control power supply and obtain the original sequence X of the grey GM (1,1) prediction model composed of the output current sampling data in the last four switching cycles. (0) , the expression is:

[0091] X (0) =[X (0) (1),X (0) (2),...,X (0) (n)] (1)

[0092] Background value sequence calculation module is used to calculate the background value sequence based on the original sequence X (0) , generate the cumulative sequence X (1) (k), where k represents the switching period, and then combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k);

[0093] Gray parameter calculation module, used for background value sequence Z obtained based on background value sequence calculation module (1) (k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b; judge whether the background value sequence Z needs to be reconstructed according to the size of the gray development coefficient a (1) (k) and recalculate the grey development coefficient a and the grey quantity b;

[0094] The first current prediction module is used to predict the gray development coefficient a, gray quantity b and the original sequence X obtained in step S3. (0) , predict the output current i(k+1) of the next switching cycle, and replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) ;

[0095] The second current prediction module is used to predict the current based on the updated original sequence X (0) , repeatedly run the background value sequence calculation module, the gray parameter calculation module and the first current prediction module to predict the output current i(k+2) of the next switching cycle again;

[0096] The control module is used to take i(k+2) as the latest predicted value of the output current; the signal corresponding to i(k+2) is subtracted from the reference signal to obtain the output current error; when the output current error is small, the proportional control parameter of the EAST fast control power supply is adjusted to be smaller; when the output current error is large, the proportional control parameter of the EAST fast control power supply is adjusted to be larger to adjust the duty cycle of the power switch tube.

[0097] The grey parameter calculation module includes a grey development coefficient a determination unit, and its specific operation mode is as follows:

[0098] When a>0, the background value sequence Z is retained (1) (k), and retain the calculated gray development coefficient a and gray amount b;

[0099] When a<0, the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), and based on the reconstructed cumulative sequence X Δ(1) (k) Recalculate the background value sequence Z (1) (k), and then based on the recalculated background value sequence Z (1) (k) Update and calculate the gray development coefficient a and gray amount b.

[0100] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. The improved grey GM (1,1) prediction method for the output current of the EAST fast-controlled power supply is characterized by: The following steps are involved: S1. Sample the output current of the EAST fast control power supply to obtain the original sequence X of the grey GM (1,1) prediction model composed of the output current sampling data in the last four switching cycles. (0) , the expression is: X (0) =[X (0) (1),X (0) (2),…,X (0) (n)] (1) S2, based on the original sequence X (0) , generate the cumulative sequence X (1) (k), where k represents the switching period, and then combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k); S3, based on the background value sequence Z obtained in step S2 (1) (k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b; judge whether the background value sequence Z needs to be reconstructed according to the size of the gray development coefficient a (1) (k) and recalculate the grey development coefficient a and the grey quantity b; S4, based on the gray development coefficient a, gray quantity b and original sequence X obtained in step S3 (0) , predict the output current i(k+1) of the next switching cycle, and replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) ; S5, based on the updated original sequence X (0) , repeat steps S2 to S4 to predict the output current i(k+2) of the next switching cycle again; S6. Take i(k+2) as the latest predicted value of the output current; subtract the signal corresponding to i(k+2) from the reference signal to obtain the output current error; when the output current error is small, adjust the proportional control parameter of the EAST fast-control power supply to be smaller; when the output current error is large, adjust the proportional control parameter of the EAST fast-control power supply to be larger to adjust the duty cycle of the power switch tube.

2. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 1 is characterized in that: The cumulative sequence expression described in step S2 is: 。 3. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 2 is characterized in that: Step S2 combines the calculus idea to accumulate the sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k) Specifically: The accumulated sequence X (1) (k) The curve is divided into l rectangles, and the area of ​​l rectangles is used to approximate the integral value of the curve. (1) (k) After segmentation, the background value sequence Z is approximately calculated (1) (k), the background value sequence expression is obtained as: Where k represents the switching period.

4. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 3 is characterized in that: The gray development coefficient a and gray amount b are calculated in step S3, and the specific calculation expressions are: X n =[X (0) (2) X (0) (3) X (0) (4)] T (5) 。 5. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 4 is characterized in that: Step S3 determines whether the background value sequence Z needs to be reconstructed by the size of the gray development coefficient a. (1) (k) and update the calculation of the gray development coefficient a and the gray amount b, specifically: When a>0, the background value sequence Z is retained (1) (k), and retain the calculated gray development coefficient a and gray amount b; When a<0, the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), and based on the reconstructed cumulative sequence X Δ(1) (k) Recalculate the background value sequence Z (1) (k), and then based on the recalculated background value sequence Z (1) (k) Update and calculate the gray development coefficient a and gray amount b.

6. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 5 is characterized in that: The accumulated sequence X is reconstructed by the transformation operator Δ(1) (k) specifically: Reconstruct the cumulative sequence X Δ(1) In the process (k), the original sequence error caused by the inconsistency of concavity and convexity is first calculated as Δ(k+1)=X (1) (n-k)-X (1) (n-k-1),k=0,1,…,n-2 (8) The reconstructed cumulative sequence is 。 7. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 6 is characterized in that: The output current i(k+1) of the next switching cycle described in step S4 is predicted by: 。 8. The improved grey GM (1,1) prediction method for output current of the EAST fast-controlled power supply according to claim 6 is characterized in that: The original sequence X updated in step S4 (0) The expression is X (0) =[X (0) (2),…,X (0) (n),i(k+1)] (11)。 9. The improved grey GM (1,1) prediction system for the output current of the EAST fast-controlled power supply is characterized by: When the system is running, the method described in any one of claims 1 to 8 is adopted, including the following modules: The sampling module is used to sample the output current of the EAST fast control power supply and obtain the original sequence X of the grey GM (1,1) prediction model composed of the output current sampling data in the last four switching cycles. (0) , the expression is: X (0) =[X (0) (1),X (0) (2),...,X (0) (n)] (1) Background value sequence calculation module is used to calculate the background value sequence based on the original sequence X (0) , generate the cumulative sequence X (1) (k), where k represents the switching period, and then combined with the idea of ​​calculus, the cumulative sequence X (1) (k) The background value sequence Z is obtained by segmentation approximation calculation (1) (k); Gray parameter calculation module, used for background value sequence Z obtained based on background value sequence calculation module (1) (k) and the original current sampling data, calculate the gray development coefficient a and the gray amount b; judge whether the background value sequence Z needs to be reconstructed according to the size of the gray development coefficient a (1) (k) and recalculate the grey development coefficient a and the grey quantity b; The first current prediction module is used to predict the gray development coefficient a, gray quantity b and the original sequence X obtained in step S3. (0) , predict the output current i(k+1) of the next switching cycle, and replace the original sequence X with the output current i(k+1) of the next switching cycle (0) The oldest current sampling value in the original sequence X is obtained (0) ; The second current prediction module is used to predict the current based on the updated original sequence X (0) , repeatedly run the background value sequence calculation module, the gray parameter calculation module and the first current prediction module to predict the output current i(k+2) of the next switching cycle again; The control module is used to use i(k+2) as the latest predicted value of the output current; the signal corresponding to i(k+2) is subtracted from the reference signal to obtain the output current error; when the output current error is small, the proportional control parameter of the EAST fast control power supply is adjusted to be smaller; when the output current error is large, the proportional control parameter of the EAST fast control power supply is adjusted to be larger to adjust the duty cycle of the power switch tube.

10. The EAST fast-controlled power supply output current improved grey GM (1,1) prediction system according to claim 9, characterized in that: The grey parameter calculation module includes a grey development coefficient a determination unit, and its specific operation mode is as follows: When a>0, the background value sequence Z is retained (1) (k), and retain the calculated gray development coefficient a and gray amount b; When a<0, the cumulative sequence X is reconstructed through the transformation operator Δ(1) (k), and based on the reconstructed cumulative sequence X Δ(1) (k) Recalculate the background value sequence Z (1) (k), and then based on the recalculated background value sequence Z (1) (k) Update and calculate the gray development coefficient a and gray amount b.

Images