Research On Mental Load Of On-duty Engineers In Nuclear Power Plant Commissioning Based On Cognitive Task Analysis

Nowadays,nuclear power has become an important source of electricity in our country.Safety is the primary task of nuclear power plants because of the nuclear power is a high safety industry And most of the operations in the main control room of the nuclear power plant are mainly cognitive and monitoring tasks.Nuclear power workers work hard,especially the commissioning engineers are often under tremendous work pressure and mental workload.In addition,with the application of new digitized technologies in nuclear power plants,a large number of information display and human-computer interaction interface have made it more difficult for commissioning engineers on duty,which easily leads to their mental overload and greatly increases the risk of human error.There is a lack of guidance and countermeasures for such tasks in the nuclear power environment where cognitive tasks are the main task,which poses a potential threat to the safe operation of nuclear power and may lead to serious accidents In this paper,in order to study the mental workload of the commissioning engineers of the nuclear power plant.We analyzed the work tasks of the nuclear power commissioning engineers,and evaluated the mental workload of the commissioning engineer on duty,in order to reveal the effects of different the commissioning task elements on the mental workload,and puts forward some effective measures and advice to reduce human error for supporting the nuclear power plant work arrangements and management decision.This study explored the mental workload changes of the commissioning engineers when they experience four task difficulty levels(the difficulty increased from 1 to 4,which are called task 1(difficulty 1),task 2(difficulty 2),task 3(difficulty 3)and task 4(difficulty 4).)respectively.Eventually,a multi-modal mental workload assessment and prediction model is established based on task analysis,and the accuracy of mental workload prediction model is verifiedIn this study,a human factor experiment was carried out,where within-subjects design was adopted,and twenty-two(N=22)the nuclear power commissioning engineers were required to complete a series of nuclear power commissioning simulation experiments.EEG(extracted ?,?,?,?,?,?/?,?+?)/? power of brain waves from the original data)and eye movement signal(blink counts,blink duration,blink rate,fixation counts,fixation duration,fixation rate,saccade counts,saccade rate,saccade duration,saccade amplitude,saccade velocity and the pupil diameter)were extracted as well as subjective mental workload score(NASA TLX scale).The experimental data were analyzed by repeated measurement analysis of variance(ANOVAs),Pearson correlation and multiple linear regression analysis,as well as GMDH neural network algorithmANOVAs analysis demonstrated that 6 kinds of EEG metrics(?,?,?,?,?,?/? brain waves power)there is no significant difference in different task difficulty,task difficulty on EEG metrics basically have no significant influence(p>0.05).The task difficulty had a significant effect on blink counts,blink duration,blink frequency,fixation counts,saccade counts and pupil diameter(p<0.05),but no significant effects on other EEG indicators(p>0.05).However,there was a significant difference in the score of subjective mental workload under different task difficulty,that is to say the task difficulty has a significant impact on subjective mental workload.Pearson correlation analysis showed that there was no existed significant linear correlation between EEG,eye movement and subjective mental workload under different task difficulty levels(p>0.05),but in general,blink duration,fixation counts,fixation frequency and saccade counts were significantly correlated with NASA-TLX scores.However,there was a significant correlation between the predicted value of mental workload and the subjective mental workload(NASA-TLX scores)based on task analysis.Regression analysis showed that?,?,?,?,?/? brain waves power and the blink duration could effectively predict the subj ective mental workload.In addition,our study in order to predict the mental workload of the nuclear power commissioning engineers at work and obtain more useful and accurate information from the original data,this study used multiple linear regression analysis(MLR)and GMDH algorithm to establish a comprehensive mental workload prediction model.The experimental results showed that both the mental workload prediction models established by the two methods can accurately predict the mental workload of commissioning engineers.Moreover,the GMDH neural network algorithm is more effective and accurate.This research verified the accuracy and reliability of the mental workload prediction methods based on task analysis,and the effect of prediction was better than the subjective NASA-TLXThe results of this study could guide the establishment of mental workload assessment and management system for commissioning engineers,and are of great significance to the prevention of nuclear power safety risks caused by excessive mental workload in advance.It provides a theoretical basis for the task design and establishment of the mental workload assessment system for the nuclear power commissioning engineers in the future,which has important guidance and reference significance,and plays an important role in improving the operation efficiency of nuclear power plants.And the results of the experiment also could be widely used in engineering application.