Abstract—

The performance efficiency of automated intelligent systems (AISs), which provide the managing personnel of electric power systems not only with systematically ordered information about the technical state of equipment and installations but also with recommendations on arranging their operation, maintenance, and repair, depends first of all on the safety and faultlessness of the relevant databases. Continuous monitoring of the indicators characterizing the technical state of equipment and installations involves high costs that are far from always being justified. Therefore, in most frequent cases, these indicators are determined from the results of tests and emergency and scheduled repairs. In fact, this information is discrete in nature and is entered in the database from dedicated logbooks. The urgency of automated settling of matters concerned with arranging maintenance and repair becomes even more important in view of the fact that no less than half of the main equipment and installations operating in electric power systems have worked out their fleet life in many respects. The use of indicators like thermal efficiency margin or permissible number of short circuit fault clearances by a circuit breaker for technical state management purposes leads to a higher risk of making erroneous decisions under these conditions. Therefore, the urgency of the problem of ensuring the safety and faultlessness of AIS databases does not decrease with time but, on the contrary, constantly tends to become more important. As an example of incorrectness of the existing approach to recognition of gross errors, the article considers data on the monthly average values of technical and economic indicators of the boiler installations of gas-and-oil fired 300-MW power units. It is pointed out that the sample of monthly average values of technical and economic indicators is inconsistent with the representative sample from the general totality of data. An interval checking method and a checksum method for recognizing gross errors have been developed and approbated. The first method is based on comparing the realizations of technical and economic indicators with their possible variation interval, and the second method is based on comparing the estimated and real annual average values of the realizations of technical and economic indicators. By using the proposed methods, it is possible to decrease the risks of elaborating erroneous recommendations, making erroneous decisions, and spending excessive costs.

中文翻译：

---

自动化分析锅炉机组技术经济指标中的数据无误

摘要-

自动化智能系统（AIS）的性能效率首先取决于电力系统的管理人员，这些自动化系统不仅向电力系统的管理人员提供有关设备和装置的技术状态的系统排序信息，而且还提供有关安排其运行，维护和维修的建议。首先是有关数据库的安全性和无误性。持续监控表征设备和装置技术状态的指标涉及高昂的成本，远远不能证明是合理的。因此，在大多数情况下，这些指标是根据测试结果以及紧急情况和定期维修确定的。实际上，该信息本质上是离散的，并且是从专用日志中输入到数据库中的。鉴于至少有一半的电力系统主要设备和装置已经在许多方面解决了其船队寿命这一事实，与维护和修理有关的事项自动解决的紧迫性变得更加重要。出于技术状态管理目的，使用诸如热效率裕度或断路器允许的短路故障间隙数量之类的指标会导致在这些条件下做出错误决定的风险更高。因此，确保AIS数据库的安全性和无故障性问题的紧迫性不会随着时间的推移而降低，相反，它往往会变得越来越重要。举例来说，现有的重大错误识别方法不正确，本文考虑了使用300兆瓦燃气和燃油发电设备的锅炉装置的技术和经济指标每月平均值的数据。需要指出的是，技术和经济指标的月平均值样本与总体数据中的代表性样本不一致。已经开发并认可了用于识别严重错误的间隔检查方法和校验和方法。第一种方法是将技术和经济指标的实现与可能的变化区间进行比较，第二种方法是将技术和经济指标的实现的估计年平均值与实际年均值进行比较。通过使用建议的方法，可以降低制定错误建议的风险，