【作者】 杨璧玲；

【导师】 晏雄；

【作者基本信息】 东华大学 ， 纺织工程， 2009， 博士

【摘要】 超高分子量聚乙烯（UHMWPE）纤维具有优异的综合性能,以其作为增强体的聚乙烯自增强（PE/PE）热塑性复合材料由于纤维和基体材料的化学相容性,可很好地发挥UHMWPE纤维的综合优良性能,因此,是一种极具应用前景的先进复合材料。研究PE/PE复合材料的损伤失效机理,对确保其结构在服役过程中的力学安全性是十分重要的,也有利于获得最佳的结构设计。现有对复合材料损伤失效机理的研究大多是集中在碳纤维、玻璃纤维等增强的热固性复合材料上。然而,由于基体的塑性性能、复合材料的热塑成型等特点,热塑性复合材料力学行为与热固性复合材料有很大的差异,而近年来随着热塑性复合材料的快速发展,工程领域应用热塑性复合材料越来越广泛,因此,对热塑性复合材料损伤失效机理的研究很有必要。本文对PE/PE热塑性复合材料损伤失效机理的研究具有现实意义。纤维增强复合材料的损伤是一个复杂的逐渐破坏过程,一般包含有基体开裂、纤维断裂、纤维抽拔、纤维-基体界面脱粘和分层等多种损伤模式。为了准确揭示PE/PE复合材料的逐渐损伤过程,本文开展了两个部分的研究工作:一是以基于逐渐损伤分析的有限元法作数值模拟研究,实现对层合板损伤进程及强度等性能的预报;二是以基于信号模式识别（PR）分析的声发射（AE）技术作实验检测研究,获取关于复合材料损伤进程的更丰富信息,并实现对模型的验证。本文为两种方法研究中尚存在的很多问题提供了新的解决思路,并以期为实际工程应用提供一种简便有效的解释热塑性复合材料损伤机制与检测分析的方法。在数值模拟研究中,首先开展了几种UHMWPE/LDPE单层板（0°、90°和45°）的力学测试实验,对实验结果初步分析发现,该材料的纵向、横向与偏轴剪切拉伸均存在明显的非线性,其纵向拉伸强度远大于其它两个方向上的拉伸强度。针对这一特性,以分段线弹性方法考虑了复合材料各向材料属性(E₁、E₂和G₁₂)的非线性,并结合复合材料层合板的逐渐损伤有限元分析方法提出了一种综合考虑纵向、横向与面内剪切非线性的PE/PE层合板逐渐损伤模型。然后,利用ANSYS软件对所建立的含材料非线性的PE/PE层合板逐渐损伤模型进行模拟,采用APDL语言编写了参数化的有限元分析程序。模拟研究了单层板的材料非线性对其力学行为的影响及层合板的逐渐损伤过程。与拉伸试验数据对比的结果表明:UHMWPE/LDPE复合材料的各向材料属性(E₁、E₂和G₁₂)非线性对于复合材料力学行为的数值模拟具有重要影响,而E₁的非线性对于复合材料层合板的拉伸力学行为有着显著的影响;以分段线弹性的方法表征各向材料属性的非线性,可简便有效地分析复合材料及其结构由于材料非线性而造成的非线性力学特性;所建立的含材料非线性的逐渐损伤模型清楚地揭示了UHMWPE/LDPE层合板在拉伸损伤过程中的基体开裂、纤维-基体脱粘和纤维断裂等损伤模式,并能反映损伤的进程,它简便地呈现了复合材料层合板的非线性拉伸过程,在应力-应变发展趋势与拉伸强度、断裂应变等性能预测上都与拉伸试验结果相一致,取得了良好的模拟结果。在声发射检测实验研究中,首先研究了几种简单铺层（0°、90°和[±45°]）UHMWPE/LDPE复合材料的损伤AE特征及失效机理。采用模式识别（PR）一般分析方法对AE数据进行了预处理和聚类分析。在聚类分析中,结合扫描电子显微镜（SEM）技术获取了较好的分类结果。结果表明,PR技术对UHMWPE/LDPE热塑性复合材料损伤AE信号的区分是客观适用的,并可有效地对信号进行除噪;PR方法能别出试样中的基体开裂、纤维-基体界面脱粘、纤维抽拔和纤维断裂等损伤模式,识别结果与利用SEM对破坏断面观察所得结果一致;借助AE信号累计数对应变的关系曲线,能清楚了解试样中各种信号类别的活动特性,再结合每种试样的特点,并可较合理地辨清损伤模式与信号类别的对应关系,从而掌握试样的损伤活动特性。因此,本研究以0°、90°和[±45°]UHMWPE/LDPE热塑性复合材料为分析对象展开了进一步的研究,辅以对LDPE基体和UHMWPE纤维的声发射检测分析,建立起一种能客观识别复合材料试样中未知信号类别损伤源机制的无监督识别（UPR）统一方法,并由此获取了UHMWPE/LDPE复合材料中不同损伤模式的AE信号典型波形特征与信号样本数据,以期为复杂铺层层合板损伤AE信号的分类与损伤源机制识别提供有效分析方法。根据这些分析方法与结果,采用UPR和有监督识别（SPR）对UHMWPE/LDPE准各向同性层合板（[0/90/+45]_s）的损伤AE信号进行了比较分析,并考察其损伤失效机理。结果表明,以所建立的UPR统一方法开展的无监督识别分析可客观、简便地分析复杂铺层层合板的损伤AE信号,另一方面,通过选取适当的样本数据,SPR方法同样可以实现对复杂铺层层合板损伤AE信号的快速区分;本文设计了从基体、纤维和简单铺层复合材料的分析中获取各种损伤模式的典型信号波形特征与AE信号数据,为上述两种分析方法在复杂铺层复合材料中的应用解决了关键性的问题,获得了合理的分析结果;两种方法对UHMWPE/LDPE层合板的研究得到了较一致的分类结果,表明PR技术分析具有客观一致性,层合板中不同损伤源机制所产的信号具有可分性。分析结果显示,对于[0/90/±45]_s层合板,其主要损伤机制有基体变形开裂、纤维断裂、和纤维-基体界面脱粘破坏,并伴有少量的纤维抽拔损伤,这些损伤都有一个发展的过程,其中纤维的损伤与断裂是层合板失效的主要原因。以这一结果对逐渐损伤模型所揭示的层合板损伤进程及机理进行了验证。研究最终表明,本文建立的含材料非线性的逐渐损伤模型可准确预报层合板的损伤特性,获得有关损伤的状态与进程;模拟的结果与声发射检测分析结果（包括SEM分析结果）相一致;同时,模型在对层合板应力-应变的发展趋势及其最终拉伸强度、断裂应变等均可以良好精度地模拟预测,验证了模型的有效性。更多还原

【Abstract】 Ultra-high Molecular Weight Polyethylene （UHMWPE） has an excellent integrated property. Self-reinforced polyethylene （PE/PE） thermoplastic composites with UHMWPE as the reinforcing fiber can transfer the excellent properties of UHMWPE fiber to the composites commendably because of the chemical compatibility of fiber and matrix. Therefore, PE/PE composites have a promising future in application. Investigations on the damage and failure mechanisms of PE/PE composites are important to the assurance of mechanical safety of the composite structures during serviced, and the obtainment of optimal design of structures. Researches hitherto mostly focus on carbon fiber and glass fiber reinforced thermoset composites. However, because of the factors such as the plasticity of matrix and the thermoplastic of the composites, there exists a big difference between thermoplastic and thermoset composites in mechanical behaviors. On the other hand, as the faster development of thermoplastic composites in the recent years, thermoplastic composites are more and more used in the engineering field. So investigations on thermoplastic composites are a need of time. There are realistic meanings in the study on the PE/PE thermoplastic composites.The damage and failure process of fiber-reinforced composites is a complicate and progressive one. It may include many damage modes such as matrix cracking, fiber breakage, fiber pullout, fiber-matrix debonding and delamination. In order to reveal the progressive damage process of PE/PE composites nicely, this investigating study carried out two parts of work: the first part was numerical modeling study based on the Progressive Failure Analysis Methodology using Finite Element Method （FEM）; the second part was Acoustic Emission （AE） monitoring and signal analysis based on Pattern Recognition （PR） technique. Efforts were also made to settle many existing problems in the two researching fields. The studies are expected to provide an effective and convenient way for the explaining of damage mechanisms and the analyzing of AE signals for thermoplastic composites. In the modeling study, several types of unidirectional UHMWPE/LDPE composites （0°, 90°and 45°） were firstly tested for the normal mechanical properties. Results showed that the composites have an obvious nonlinearity in longitudinal, transverse and shear stress-strain behavior. Therefore, a FEM based progressive damage model was developed for the PE/PE laminates subjected to tensile loading, by considering the nonlinear properties of unidirectional composites in the three tensile modulus(E₁, E₂ and G₁₂) as piecewise linear-elastic properties.Then, ANSYS software was used for simulating the progressive failure of UHMWPE/LDPE laminates to study the damage mechanisms, according to the established progressive failure model. The FEM analysis program was developed by using the APDL program language. Comparison was made to the data obtained from the tensile tests. Results showed as follows: the nonlinear properties of the compositesin the three tensile modulus (E₁,E₂ and G₁₂) have an important effect on the numericalmodeling of mechanical behaviors of UHMWPE/LDPE laminates; the nonlinear behavior of the laminates were remarkably affected by the longitudinal nonlinearity of the composites; the nonlinearity of composites and structures was conveniently and effectively included by the piecewise linear elasticity treatment; the progressive failure model can predict damage modes such as matrix cracking, fiber-matrix shearing to debonding and fiber fracture, and reveal the damage propagation; the analytical prediction showed an excellent agreement with the experimental data in the tensile stress, rupture strain and the developing trend of the stress-strain curve.In the AE monitoring and study, several simple lay-up of UHMWPE/LDPE composites （0°, 90°and [±45°]） were firstly investigated for the AE features and damage mechanisms of the composites. A common analytical procedure of PR technique was used for the preprocessing and clustering analysis of the AE data. In the clustering analysis, scanning electron microscope （SEM） technique was utilized for the assuring of the classification. Results showed as follows: the PR technique is objective and suitable for the analysis of AE data from UHMWPE/LDPE thermoplastic composites; it also can perform noise reduction effectively; the PR technique is able to identify damage modes such as matrix cracking, fiber-matrix debonding, fiber pullout, fiber breakage, etc. in the specimens, and the identification results are the same with the observation results by SEM; by the cumulative AE hits of each damage mode vs. strain curves, the damage process of the specimens can be reviewed clearly; in addition to the features of the specimen, a correlation between the clustered AE signal classes and their original damage modes could be established reasonably, and a clear understanding of damage mechanisms in the composites could be finally reached.Therefore, investigations were further carried out for the simple lay-up laminate of 0°, 90°and [±45°] UHMWPE/LDPE composites. Fracture waveforms of pure PE resin and UHMWPE fiber bundle were collected analyzed, in order to establish, by using Unsupervised Pattern Recognition （UPR） technique, a unified analytical procedure which can recognize the mechanical souses of the clustered AE signal classes objectively. Typical waveforms and AE signal samples of different damage modes in UHMWPE/LDPE composites were obtained from the analysis. All the information and the unified analytical procedure using UPR technique were proposed to be assistant in the analysis of the cases of complicated lay-up laminates. Then, AE data from UHMWPE/LDPE quasi-isotropic laminates （[0/90/±45]_s） was analyzed to study the damage mechanisms, by compared analysis using the established UPR procedure and Supervised Pattern Recognition （SPR） technique respectively. Results showed that the established UPR procedure can recognize the mechanical sources in the complicated lay-up laminates effectively; that on the other hand, by selecting a suitable sample data, the SPR technique can reach a fast separation of AE data from complicated lay-up laminates. In this study, typical waveforms and AE signal samples of different damage modes in UHMWPE/ LDPE composites were set to be obtained from the analyses of matrix, fiber and simple lay-up laminates. By this way the critical problems in the use of the UPR and SPR technique for the AE classification were solved and reasonable classifying results were obtained. Classifying results by the two techniques reached a good coherence. It showed that the PR technique is objective to the AE data analysis of UHMWPE/ LDPE composites, and that AE signals of the composite laminates are of separableness.The analytical result showed that, as for the [0/90/±45]_s laminates, the main damage modes existing were matrix cracking, fiber breakage and fiber-matrix debonding, with a little fiber pullout. All these damage modes presented a progressive failure process. The damage and failure of fibers was the dominant failure mode in the laminates and it account for the final fracture of the laminates. These results showed by the AE technique were used for the validation of the established progressive failure model. The investigating studies finally showed that, the established progressive failure model, considering the material nonlineanty of the UHMWPE/LDPE composites, can correctly predict the failure process and reveal the damage mechanisms. The numerical modeling results were supported by the AE monitoring and analyzing （including the SEM analyzing） results. At the same time, the model can predict the tensile stress, rupture strain and the developing trend of the stress-strain curve for the laminates perfectly, proving that the numerical modeling is valid.更多还原