Synthesis ammonia first-section furnace operation anomaly detection method based on time sequence difference feature extraction

Abstract

The invention discloses a synthetic ammonia first-section furnace operation anomaly detection method based on time sequence difference feature extraction, and aims to adaptively extract time sequence difference features from synthetic ammonia first-section furnace process data in real time so as to realize and complete a high-sensitivity synthetic ammonia first-section furnace operation anomaly detection task. Specifically, according to the method, a time sequence error generation model is built in a self-adaptive mode through process data of a synthesis ammonia first-section furnace. And on the basis, the operation anomaly detection of the synthesis ammonia first-section furnace is implemented by using the error of the online autoregression model. The method has the biggest technical advantages that a brand new time sequence difference characteristic analysis algorithm technology is adopted to perform characteristic analysis and extraction on the process data of the synthesis ammonia first-section furnace, and the time sequence difference characteristic of the maximum time sequence error can be obtained through self-adaptive updating. Therefore, from this perspective, the method of the invention can theoretically use high-sensitivity time sequence difference characteristics to implement anomaly detection.

Description

technical field [0001] The invention relates to a method for detecting abnormality in a chemical process, in particular to a method for detecting abnormal operation of a first-stage synthetic ammonia furnace based on time series difference feature extraction. Background technique [0002] Ammonia is one of the important inorganic chemical products and occupies an important position in the national economy. The synthetic ammonia production process compresses the raw materials such as coke, coal, coke oven gas, natural gas, naphtha, heavy oil and feed gas (semi-water gas) to a certain pressure after purification and refining (the main components are hydrogen and nitrogen), and then enters the ammonia synthesis tower. , Ammonia and other by-products are generated under the action of high temperature and catalyst. The first-stage furnace is the key equipment of the synthetic ammonia plant in the fertilizer plant. The first-stage ammonia-synthesis furnace transforms natural gas ...

AI Technical Summary

Problems solved by technology

Traditional time series feature analysis and extraction techniques that can be used for anomaly detection usually only focus on the mining of time series relationship features under normal or expected operating conditions. Whether the mined time series relationship features are beneficial to anomaly detection is unknown.

Images