核电结构材料应力腐蚀开裂是影响核电设备长期安全运行的关键问题之一。在模拟核电结构材料实际工作条件的高压釜环境下进行裂纹扩展速率监测试验是核电站安全评价和设备寿命预测工作的重要组成部分。鉴于高压釜环境的特殊性，直流电位降（DCPD）法是能够应用于高压釜环境下监测裂纹扩展的少数手段之一。实际应用中，表征裂纹长度的电位降信号不但受裂纹扩展影响，还受多种非裂纹扩展因素影响，且信号微弱易被干扰。为了厘清影响电位降信号的因素，提高裂纹扩展信号的监测精度，本文对直流电位降法裂纹扩展监测的机理和微弱电位降信号抗干扰测量方法展开研究，主要研究工作和研究结果如下：

研究了典型含裂纹试样的电流密度分布和电位场分布，分析了影响电位降与裂纹长度关系的主要因素和二者的归一化关系。通过含裂纹试样电位场分布求解，研究了裂纹实际形态特征对测量结果的影响。结果表明，裂纹前沿弯曲使裂纹扩展测量值低于真实值，而裂纹偏斜、锯齿形裂纹和裂纹分叉则使测量值高于真实值。利用实际裂纹形态特征建模进行测量值修正可以提高裂纹扩展测量精度。

研究了试样接线点布局对主电位降信号灵敏度、幅度和测量点位置误差敏感度的影响，以及对参比电位降信号可测性和裂纹相关度的影响。提出了参比电位降测量点布局的选择依据和基于多目标遗传算法的优化方法。优化得到的测量点位置的Pareto解均位于试样边界以内，与位于试样边界的情况相比，在保证低裂纹相关度的同时可使信号幅度提高1个数量级。

根据直流电位降信号监测实验数据的时域分析、频域分析及统计特征参量分析，研究了直流电位降信号及干扰信号的特征，并据此建立了直流电位降信号模型。研究结果表明，时变热电势、温度相关干扰和随机噪声是影响微弱直流电位降信号的主要干扰因素。以上干扰信号特征不同，热电势信号和随机噪声信号表现为加性，而温度相关干扰信号则表现为乘性。热电势信号大小与试样材料、导线材料、温度均有关，高压釜环境下热电势信号比空气环境下大1个数量级，并具有比较明显的周期性。

根据建立的直流电位降信号模型，对模型包含的三种干扰信号分别设计信号处理算法，提出了联合小波阈值去噪与变电流信号合成技术的信号抗干扰处理算法。信号中的热电势干扰信号和噪声信号能够被有效抑制，同时乘性干扰信号则被保留，从而能够进一步使用参比电位降法予以抑制。对空气环境和高压釜环境下的直流电位降信号监测实验数据的处理效果表明，提出的信号处理算法可以有效抑制干扰，使信噪比提高约30分贝。

通过在试样慢拉伸实验和低频疲劳拉伸实验过程中实时监测电位降信号，结合试样形变及电场仿真分析，研究了试样受载变形对电位降的影响。实验测得了与载荷频率一致的电位降信号波动并分析其变化规律，波动原因包括裂纹尖端附近的裂纹面接触和电阻应变效应两方面。这种载荷变化致电位降信号变化使裂纹扩展测量产生误差，对于一定幅度的载荷，测量误差会随着裂纹的扩展越来越大。在此基础上将材料和载荷参数引入标定函数，提出了纳入载荷影响的标定方法，能够校正因载荷变化而产生的测量误差。

研制了集成化裂纹扩展监测仪样机，实现了微伏级信号测量。以其为信号监测实验平台，进行了高压釜环境下紧凑拉伸试样的裂纹扩展信号监测实验，采用优化后的接线点布局，使用提出的信号抗干扰处理算法抑制干扰信号，并按照纳入载荷影响的标定方法将直流电位降信号转换为裂纹扩展信号，最终使裂纹扩展信号标准差小至5微米。

The stress corrosion cracking of structural materials in nuclear power plants is one of the critical problems affecting the long-term safe operation of nuclear power equipment. The monitoring test of crack growth rate in the autoclave environment that simulates the actual working conditions of structural materials in nuclear power plants is an essential part of the safety evaluation and equipment life prediction of nuclear power plants. Given the exceptional environment, the direct current potential drop (DCPD) method is one of the few methods that can be applied to monitor crack propagation in the autoclave environment. In practice, the potential drop signal representing the crack length is not only affected by crack propagation but also by various non-crack propagation factors, and the signal is weak and easily disturbed. To clarify the factors affecting the potential drop signal and to improve the accuracy of crack propagation signal monitoring, the mechanism of monitoring crack propagation by direct current potential drop method and the anti-interference measurement of weak potential drop signal were studied in this dissertation. The main research work and results are as follows:

The current density distribution and potential field distribution of typical cracked specimens were studied. The main factors affecting the relationship between potential drop and crack length and their normalized relationship were analyzed. By solving the potential field distribution of the specimen with cracks, the influence of crack morphology on measurement was analyzed. The results have shown that crack front bending causes crack growth measurements to be lower than true, while crack deflection, zigzag cracks, and crack bifurcation cause measurements to be higher than true. The accuracy of crack propagation measurement can be improved by using the actual crack morphology feature modeling to correct the measured values.

The influence of the wiring points layout on the sensitivity, the amplitude, and the measurement position error sensitivity of the main potential drop signal, as well as the influence on the measurability and the crack correlation of the reference potential drop signal, were studied. The selection basis of measurement points layout of reference potential drop and the optimization method based on the multi-objective genetic algorithm were proposed. The optimized Pareto solutions of the measurement points positions are all located within the specimen boundary. Compared with the case where the measurement points are at the boundary of specimens, the signal amplitude can be increased by order of magnitude while ensuring low crack correlation.

According to the time-domain analysis, frequency-domain analysis and statistical characteristic parameter analysis of the experimental data of the direct current potential drop signal monitoring, the characteristics of the signal and the interference signal were studied, and the direct current potential drop signal model has been proposed accordingly. The results show that the time-varying thermal electromotive force, temperature-related interference, and random noise are the main interference factors affecting the weak direct current potential drop signal. The characteristics of the above interference signals are different. The thermal electromotive force and random noise appear additive, while the temperature-dependent interference signal appears multiplicative. The thermal electromotive force is related to the specimen material, wire material, and temperature. The thermal electromotive force in an autoclave environment is one order of magnitude larger than that in air and has evident periodicity.

According to the proposed direct current potential drop signal model, signal anti-interference algorithms were designed for the three interference signals included in the model. A signal anti-interference processing algorithm combining wavelet threshold denoising and variable current signal synthesis technique was proposed. The thermal electromotive force signal and noise signal can be effectively suppressed while the multiplicative interference signal is retained so that the reference potential drop method can be further used to suppress it. The processing results of the experimental data of direct current potential drop signal monitoring in air and autoclave environment show that the proposed signal processing algorithms can effectively suppress the interference and increase the signal-to-noise ratio by about 30 dB.

By monitoring the potential drop signal in real-time during the slow tensile test and low-frequency tensile fatigue test, as well as the specimen deformation and electric field simulation analysis, the influence of the loading deformation on the potential drop was studied. The fluctuation of the potential drop signal consistent with the load frequency was measured experimentally, and its variation law was analyzed. The reasons for the fluctuation are the crack surface contact near the crack tip and the resistance strain effect. The change of potential drop signal induced by the load causes errors in crack propagation measurements. For a certain magnitude of load, the measurement error will increase as the crack grows. The material and load parameters were introduced into the calibration function. A calibration method incorporating the influence of the load was proposed, which can correct the measurement error caused by the load change.

A prototype of an integrated crack growth monitor was developed, and the signal measurement at the microvolt level was realized. The crack propagation signal monitoring experiment of the compact tensile specimen in an autoclave environment was carried out using the prototype. The optimized wiring points layout was adopted. The proposed signal anti-interference processing algorithm was used to suppress the interference signal. And the direct current potential drop signal was converted into crack propagation signal according to the calibration method incorporating the influence of load. The standard deviation of the crack propagation signal was reduced to 5 μm.