论文部分内容阅读
作者提出了新的强度判据和正负偏差判据。结果表明,上述判据可有效地防止低含量相的漏检,并可解决固溶体相的漏检与误检问题。此外,还引入了人工智能方法。实践证明,它能识别其它判据无法识别的各种判据的匹配率都较高的误检相,以及同晶型、结构相似的物相,使程序给出了肯定性结果,指出可能存在相,把误检相减少到较低限度。文中对这些方法及其结果作了较详细的讨论。
The authors propose new strength criteria and positive and negative bias criteria. The results show that the above criterion can effectively prevent the detection of low content phase, and can solve the problem of missing detection and false detection of solid solution phase. In addition, an artificial intelligence approach was introduced. Practice has proved that it can identify the misjudgment phase with high matching rate of all kinds of criteria that other criteria can not be identified, and isomorphism, structure similar to the phase, the program gives affirmative results, pointing out that there may exist Phase, the false phase reduced to a lower limit. The paper discusses these methods and their results in more detail.