CN104810847A

CN104810847A - Commutation failure prevention method based on direct current fuzzy predictive control

Info

Publication number: CN104810847A
Application number: CN201510095247.0A
Authority: CN
Inventors: 孙国强; 高楷; 卫志农; 孙永辉; 楚云飞; 李逸驰
Original assignee: Hohai University HHU
Current assignee: Hohai University HHU
Priority date: 2015-03-03
Filing date: 2015-03-03
Publication date: 2015-07-29
Anticipated expiration: 2035-03-03
Also published as: CN104810847B

Abstract

The invention discloses a commutation failure prevention method based on direct current fuzzy predictive control. In view of a direct current predictive control method in which limitation exists in a parameter k random value in a direct current predication setting valve calculation formula and direct current at a rectifier side is reduced for a long time, limitation of slow current recovery of the direct current system is caused, a fuzzy controller is adopted to control the direct current predication setting valve, a direct current fuzzy predictive control method is brought forward, a control time duration T selected via a simulation result is set, and the direct current at the rectifier side is only reduced with a time duration T after a CFPREV module detects failure happens to an alternating current system. Through modeling and simulation, compared with the prior direct current predictive control method, effectiveness of restraining the commutation failure ability by the HVDC (high-voltage direct current) system is verified to be enhanced.

Description

A kind of commutation failure prevention method based on direct current Fuzzy Predictive Control

Technical field

The present invention relates to a kind of commutation failure prevention method based on direct current Fuzzy Predictive Control, belong to operation and control of electric power system technical field.

Background technology

HVDC (High Voltage Direct Current) transmission system (high-voltage direct current, HVDC) transmission capacity is large, loss is little, the life-span is long, fed distance is unrestricted, there is not the stable problem of AC transmission system simultaneously, be thus widely used in, submarine cable power transmission interconnected at long-distance and large-capacity power transmission, unsynchronized networks etc.Along with the construction of expanding economy and electrical network, China become DC power transmission line in the world at most, the maximum country of direct current transmission capacity.The extensive AC and DC power system formed thus, its complexity and difficulties run is rare in the world.

One of modal fault of HVDC system is exactly commutation failure.Commutation failure is unfavorable for the safe and stable operation of HVDC system, can cause the interruption of system instantaneous transfer power, and the rapid increase of direct current also can shorten the converter valve life-span.When commutation failure occurs, if the control measure taked are proper, can recover voluntarily after a cycle, system just can continue to run, otherwise will cause continuous print commutation failure, finally causes DC system locking.Therefore, the genesis mechanism of research commutation failure and braking measure thereof, significant.

Direct current forecast Control Algorithm is current a kind of comparatively advanced method of preventing commutation failure, but in the direct current of the method prediction setting value computing formula there is the limitation of random value in parameter k, and reduce rectification side direct current for a long time due to the method, direct current system electric current is caused to recover slowly, therefore, the method is further improved.

Summary of the invention

Goal of the invention: in order to solve the problem, the invention provides a kind of commutation failure prevention method based on direct current Fuzzy Predictive Control, the method is by fuzzy Control direct current prediction setting value, and arrange one and control duration T, only CFPREV module detect length after fault in ac transmission system occurs be T time in reduce rectification side direct current.Finally by emulation testing, demonstrate the method and compare the validity that direct current forecast Control Algorithm promotes HVDC system suppression commutation failure ability.

Technical scheme: a kind of commutation failure prevention method based on direct current Fuzzy Predictive Control, said method comprising the steps of:

In step 1:PSCAD, whether CFPREV module detection AC system breaks down, and breaks down, send Start signal enable direct current Fuzzy Predictive Control module if detect; Otherwise system attonity.

Step 2: the direct current Fuzzy Predictive Control module be enabled measures changes delta E and the rate of change d Δ E/dt of current inverter side AC system three-phase voltage effective value E, and is sent into the fuzzy controller in MATLAB/Simulink by the interface between PSCAD and MATLAB/Simulink.

Fuzzy controller in step 3:MATLAB/Simulink draws rectification side direct current decreasing value Δ Ic by step operation such as obfuscation, fuzzy reasoning, defuzzification, superposing control durations, and returns to PSCAD.

VDCL in step 4:PSCAD deducts the value after Δ Ic as rectification side direct current setting value Ido_rec, finally applies to rectification side direct current and controls, repeat step 1 subsequently.

Beneficial effect of the present invention is:

1. using fuzzy controller to replace the direct current prediction setting value computing formula of existing direct current forecast Control Algorithm, there is the limitation of certain randomness in the parameter k overcoming this computing formula.

2. arrange and control duration T, only CFPREV module detect length after fault in ac transmission system occurs be T time in reduce rectification side direct current, overcome former direct current forecast Control Algorithm and reduce rectification side direct current for a long time, thus cause direct current system electric current to recover limitation slowly.

3. compare direct current forecast Control Algorithm, the commutation failure prevention method based on direct current Fuzzy Predictive Control of the present invention effectively can promote HVDC system and suppress commutation failure ability.

Accompanying drawing explanation

Fig. 1: the existing CIGRE HVDC model-controlled system structured flowchart containing direct current forecast Control Algorithm;

Fig. 2: containing the CIGRE HVDC model-controlled system structured flowchart of the inventive method after improving;

Fig. 3: direct current Fuzzy Predictive Control inside modules structured flowchart;

Fig. 4: the flow chart of the inventive method;

Fig. 5: during single phase ground fault, this method is to the improvement of commutation failure;

Fig. 6: during three-phase fault, this method is to the improvement of commutation failure.

Embodiment

Below in conjunction with specific embodiment, illustrate the present invention further, these embodiments should be understood only be not used in for illustration of the present invention and limit the scope of the invention, after having read the present invention, the amendment of those skilled in the art to the various equivalent form of value of the present invention has all fallen within the application's claims limited range.

Fig. 1 is the existing CIGRE HVDC model-controlled system structured flowchart containing direct current forecast Control Algorithm, and Fig. 2 is the CIGRE HVDC model-controlled system structured flowchart containing the inventive method after improving.Direct current PREDICTIVE CONTROL module in Fig. 1 dotted line frame and follow-up MIN select module to have much room for improvement, the present invention's direct current Fuzzy Predictive Control module substitutes original direct current PREDICTIVE CONTROL module, replace former MIN to select module by signed magnitude arithmetic(al) module, export I _{do_rec}signal.

Fig. 3 is the internal structure block diagram of direct current Fuzzy Predictive Control module.The input variable of fuzzy controller is changes delta E and the rate of change d Δ E/dt of inverter side AC system three-phase voltage effective value E, and the output variable of fuzzy controller is rectification side direct current decreasing value Δ I _c.Start is the triggering signal of this system, and when Start is direct impulse, system brings into operation, and gathers union immediately and goes out current Δ E and d Δ E/dt, through the output of process direct current decreasing value Δ I such as obfuscation, fuzzy reasoning and defuzzification.Because Δ I cannot auto zero, along with alleviating of fault, long-time reduction rectification side direct current is also unfavorable for the fast quick-recovery of system, and the time of the reducing rectification side direct current too short control performance that will affect, thus, the present invention is provided with and controls duration T, only within the time that the rear length of Start triggering is T, reduces rectifier DC electric current.Herein by the analysis to test result under different T, control duration T is set to 5ms, and test result is as shown in table 5.

Fig. 4 is the flow chart of the inventive method.Concrete steps are as follows:

Fuzzy controller in step 3:MATLAB/Simulink draws rectification side direct current decreasing value Δ I by step operation such as obfuscation, fuzzy reasoning, defuzzification, superposing control durations _c, and return to PSCAD.

VDCL in step 4:PSCAD deducts Δ I _cafter value as rectification side direct current setting value I _{do_rec}, finally apply to rectification side direct current and control, repeat step 1 subsequently.

Table 1 is the list of Δ E membership function, table 2 is the list of d Δ E/dt membership function, in membership function list, the leftmost side one is classified as input variable via the numerical value after discretization, some states of the top one behavior fuzzy variable, these states are respectively: in negative large, negative, negative little, zero, just little, center, honest, corresponding abbreviation is respectively NB, NM, NS, Z, PS, PM, PB, the numerical value in table between other 0-1 is the degree of membership weight that input variable centrifugal pump is under the jurisdiction of fuzzy variable state.Obfuscation is process input variable being converted to fuzzy variable, and fuzzification process of the present invention is carried out based on table 1 and table 2.

Table 1

Table 2

Input variable Δ E and d Δ E/dt is fuzzy, and to change into the process of fuzzy variable DltE and Diff Δ E as follows: first, the value discretization of Δ E and d Δ E/dt, the such as excursion of an input variable value is-e to e, and the number range after discretization is-n to n, then its quantizing factor is

K_{e} = \frac{n}{e}

So input variable e _inumerical value after discretization is determined by formula below

n_{i} = \{\begin{matrix} - n, & k_{e} \cdot e_{i} < - n \\ n, & k_{e} \cdot e_{i} > n \\ l, & l \leq k_{e} \cdot e_{i} \leq l + 1 / 2 \\ l + 1, & l + 1 / 2 \leq k_{e} \cdot e_{i} \leq l + 1 \end{matrix}

Centrifugal pump n is found in the membership function list of the table of comparisons 1 and table 2 again _icorresponding fuzzy variable, changes into fuzzy variable DltE and Diff Δ E by input Δ E and d Δ E/dt.Such as, if the n after Δ E discretization _i=-6, then corresponding fuzzy subset DltE should be NB.

Table 3 is fuzzy control rule table.Fuzzy variable DltE and Diff Δ E is according to the fuzzy inference rule of table 3, reasoning obtains exporting fuzzy variable DltI, table 3 describes fuzzy rule in 28, these 28 kinds of fuzzy rules can determine one total, use mamdani fuzzy reasoning method to obtain determining to export fuzzy variable DltI herein, inference formula is as follows:

Wherein, DltE and Diff Δ E is the fuzzy variable that two input variables are obtained by obfuscation, and T is transposition symbol, it is composite operator.

Table 3

Table 4 is the list of DltI membership function.By defuzzification process, change into a concrete numerical value z by exporting fuzzy variable ^*, defuzzification formula of the present invention is as follows

z^{*} = \frac{Σ_{i = 0}^{n} μ_{c} (z_{i}) \cdot z_{i}}{Σ_{i = 0}^{n} μ_{c} (z_{i})}

Wherein z _ii-th element of DltI, μ _c(z _i) be z in DltI _icorresponding is subordinate to angle value, exports the membership function of fuzzy variable DltI in table 4.

Table 4

Table 5 is the critical commutation failure reactance that different T is corresponding.For determining suitable control duration T, do following test herein: a certain fault moment of Stochastic choice and fault duration are (for each module of guarantee system starts completely, choose fault moment 16s more rearward herein, fault duration 50ms), under single-phase earthing and three-phase fault two kinds of fault types, test the critical commutation failure reactance of system during different T respectively.Owing to controlling, duration T is maximum should be not more than fault duration 50ms, and thus its span is 1-50ms.Test result is as shown in table 5.Known by table 5, within the scope of 1-50ms, when T is greater than 5ms, the critical commutation failure reactance of single phase ground fault no longer reduces, and when T is greater than 3ms, the critical commutation failure reactance of three-phase fault no longer reduces, and therefore the best value of T is 5ms.

Table 5

Introduce one embodiment of the present of invention below:

With international conference on large HV electric systems direct current transportation standard test system (CIGRE HVDC Benchmark Model) for research model, add direct current prediction module and direct current fuzzy prediction module of the present invention respectively, test commutation failure situation under two kinds of disparate modules effects, test result is shown in Fig. 5 and Fig. 6, and fault type is respectively single phase ground fault and three-phase fault.

Comprehensive above-described embodiment result is known, compares original direct current forecast Control Algorithm, and the commutation failure prevention method that the present invention is based on direct current Fuzzy Predictive Control significantly improves the suppression commutation failure ability of HVDC system.

Claims

1., based on a commutation failure prevention method for direct current Fuzzy Predictive Control, it is characterized in that described method realizes successively according to the following steps:

In step 1:PSCAD, whether CFPREV module detection AC system breaks down, and breaks down, send Start signal enable direct current Fuzzy Predictive Control module if detect; Otherwise system attonity;

Step 2: the direct current Fuzzy Predictive Control module be enabled measures changes delta E and the rate of change d Δ E/dt of current inverter side AC system three-phase voltage effective value E, and is sent into the fuzzy controller in MATLAB/Simulink by the interface between PSCAD and MATLAB/Simulink;

2., as claimed in claim 1 based on the commutation failure prevention method of direct current Fuzzy Predictive Control, it is characterized in that: obfuscation is process input variable being converted to fuzzy variable, fuzzification process is carried out based on membership function; Input variable Δ E and d Δ E/dt is fuzzy, and to change into the process of fuzzy variable DltE and Diff Δ E as follows: first, the value discretization of Δ E and d Δ E/dt, the such as excursion of an input variable value is-e to e, and the number range after discretization is-n to n, then its quantizing factor is

K_{e} = \frac{n}{e}

n_{i} = \{\begin{matrix} - n, & k_{e} \cdot e_{i} < - n \\ n, & k_{e} \cdot e_{i} > n \\ l, & l \leq k_{e} \cdot e_{i} \leq l + 1 / 2 \\ l + 1, & l + 1 / 2 \leq k_{e} \cdot e_{i} \leq l + 1 \end{matrix}

Contrast membership function list again and find centrifugal pump n _icorresponding fuzzy variable, changes into fuzzy variable DltE and Diff Δ E by input Δ E and d Δ E/dt.

3. as claimed in claim 2 based on the commutation failure prevention method of direct current Fuzzy Predictive Control, it is characterized in that: fuzzy variable DltE and Diff Δ E is according to fuzzy inference rule, reasoning obtains exporting fuzzy variable DltI, use mamdani fuzzy reasoning method to obtain determining to export fuzzy variable DltI, inference formula is as follows:

Wherein, DltE and Diff Δ E is the fuzzy variable that two input variables are obtained by obfuscation, and T is transposition symbol, and o is composite operator, and R is regular matrix.