电网技术
電網技術
전망기술
POWER SYSTEM TECHNOLOGY
2012年
2期
250-256
,共7页
小波变换%模糊逻辑%特征提取%混合模糊Petri网%电力系统
小波變換%模糊邏輯%特徵提取%混閤模糊Petri網%電力繫統
소파변환%모호라집%특정제취%혼합모호Petri망%전력계통
wavelet transform%fuzzy logic%feature extraction%hybrid fuzzy Petri net%power system
提出一种基于混合模糊Petri网(hybrid fuzzy Petrinets,HFPN)的电力系统故障暂态信号识别新方法。该方法将小波变换特征提取、模糊逻辑和模糊Petri网相结合构成混合模糊Petri网,有效解决了单一的模糊Petri网无法识别故障暂态信号的缺陷,也改善了原有识别方法推理过程不容易被人理解的不足。对故障发生后1/4个周期内的三相电流和零序电流,应用小波变换提取小波能量,再经过模糊推理系统得到模糊值作为特征量,最后应用模糊Petri网进行识别。大量PSCAD/EMTDC仿真试验结果表明:该故障识别方法能快速准确地识别各类故障暂态信号,识别速度快,并且以概率的形式给出发生各种故障的可能性;基于HFPN的图形化表示方法清晰直观,不受故障时刻、过渡电阻、故障位置等因素的影响,对噪声信号和不同线路都具有较好的适应性。
提齣一種基于混閤模糊Petri網(hybrid fuzzy Petrinets,HFPN)的電力繫統故障暫態信號識彆新方法。該方法將小波變換特徵提取、模糊邏輯和模糊Petri網相結閤構成混閤模糊Petri網,有效解決瞭單一的模糊Petri網無法識彆故障暫態信號的缺陷,也改善瞭原有識彆方法推理過程不容易被人理解的不足。對故障髮生後1/4箇週期內的三相電流和零序電流,應用小波變換提取小波能量,再經過模糊推理繫統得到模糊值作為特徵量,最後應用模糊Petri網進行識彆。大量PSCAD/EMTDC倣真試驗結果錶明:該故障識彆方法能快速準確地識彆各類故障暫態信號,識彆速度快,併且以概率的形式給齣髮生各種故障的可能性;基于HFPN的圖形化錶示方法清晰直觀,不受故障時刻、過渡電阻、故障位置等因素的影響,對譟聲信號和不同線路都具有較好的適應性。
제출일충기우혼합모호Petri망(hybrid fuzzy Petrinets,HFPN)적전력계통고장잠태신호식별신방법。해방법장소파변환특정제취、모호라집화모호Petri망상결합구성혼합모호Petri망,유효해결료단일적모호Petri망무법식별고장잠태신호적결함,야개선료원유식별방법추리과정불용역피인리해적불족。대고장발생후1/4개주기내적삼상전류화령서전류,응용소파변환제취소파능량,재경과모호추리계통득도모호치작위특정량,최후응용모호Petri망진행식별。대량PSCAD/EMTDC방진시험결과표명:해고장식별방법능쾌속준학지식별각류고장잠태신호,식별속도쾌,병차이개솔적형식급출발생각충고장적가능성;기우HFPN적도형화표시방법청석직관,불수고장시각、과도전조、고장위치등인소적영향,대조성신호화불동선로도구유교호적괄응성。
A new hybrid fuzzy Petri net (HFPN) based method to recognize fault transient signals in power system is proposed. In the proposed method, the feature extraction of wavelet transform and fuzzy logic are integrated with fuzzy Petri net to constitute a HFPN to effectively solve the problem that a single fuzzy Petri net cannot recognize fault transient signals, and the defect of existing recognizing methods that the reasoning process is not easy to be understood is remedied. Using wavelet transform, the wavelet energy is extracted from three-phase current and zero-sequence current within the quarter period after the occurrence of the fault; then through fuzzy reasoning the obtained fuzzy value is taken as the characteristic quantity; finally the recognition is performed by fuzzy Petri net. Results from a lot of PSCAD/EMTDC based simulation show that the proposed fault classification method can recognize various fault transient signals rapidly and accurately, and the occurrence possibilities of various faults are given in the form of probability. The HFPN based graphical representation is clear and intuitionistic and is not affected by the factors such as moment when the fault occurs, transition resistance and fault position, thus it possesses good adaptability to the signals containing noise and various transmission lines.