Human factor engineering evaluation method and model

Abstract

The invention provides a human factor engineering evaluation method and model, and the method comprises the following steps: constructing a human factor engineering evaluation model, and setting a confidence coefficient threshold value in the human factor engineering evaluation training model; determining a to-be-evaluated operation scene, obtaining a feature data set of the operation scene, and inputting the feature data set into the human factor engineering evaluation training model; the human factor engineering evaluation training model evaluates human factor engineering of the operation scene to obtain a human factor engineering evaluation result and corresponding confidence in the operation scene; comparing the confidence with a confidence threshold, and when the confidence is greater than or equal to the confidence threshold, outputting a human factor engineering evaluation result; and when the confidence coefficient is smaller than a confidence coefficient threshold value, taking the operation scene as a new typical operation scene, and reconstructing a human factor evaluation training model. According to the human factor engineering evaluation method provided by the invention, the evaluation efficiency and the evaluation accuracy of human factor engineering are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of overall ship design, and in particular to a human factors engineering evaluation method and model. Background Art

[0002] The shipbuilding industry is a typical labor-intensive industry, which requires a large amount of human resources in the production process. However, compared with other modern manufacturing industries, the level of automation in shipbuilding is still relatively low, which makes labor occupy a more important position in the shipbuilding process. At the same time, due to the complexity of the ship structure and the particularity of the construction process, construction workers often need to work in a small and confined space and maintain an uncomfortable posture for a long time, which poses a serious threat to their health. Therefore, how to ensure the safety and health of construction workers and avoid the occurrence of occupational diseases has become an important issue that shipbuilding companies must face.

[0003] Human factors engineering assessment plays a vital role in the shipbuilding process. It can help companies identify and improve adverse factors for construction workers in the production process, thereby improving production efficiency and ensuring the safety of construction workers. However, traditional human factors engineering assessment methods are mainly based on simulation software, such as RULA. Although these methods can simulate the construction process to a certain extent and evaluate the workload and comfort of construction workers, they are overly dependent on theoretical judgments and the evaluation process is cumbersome, resulting in low accuracy in practical applications. In addition, traditional methods often ignore the individual differences and subjective feelings of construction workers, making it difficult for the evaluation results to fully reflect the actual situation. Summary of the invention

[0004] In view of the problems existing in the human factors engineering assessment in the prior art described above, the present application provides, on the one hand, a method for constructing a human factors engineering assessment training model, comprising the following steps:

[0005] Construct a typical operation scenario set Pa;

[0006] Select any typical operation scene Pb, extract the key features tbi of the typical operation scene, and construct a feature data set Tb = {tb1, tb2...tbn};

[0007] Arrange construction personnel to perform actual operations in the real environment of the typical operation scene, and perform multi-dimensional scoring on the operation scene Pb according to the feature data set to construct an initial score set Sb = {sb1, sb2...sbk};

[0008] Obtain evaluation scores: construct a weight set Wb = {wb1, wb2...wbk}, and perform weighted scoring to obtain evaluation scores Ab = sb1*wb1+sb2*wb2...sbk*wbk;

[0009] Aggregate the feature data set Tb to form a feature data set combination TT and an evaluation data set combination AA corresponding to the feature data set combination TT, TT = {T1, T2 ... Ta}, AA = {A1, A2 ... Aa};

[0010] Among them, a is a positive integer greater than or equal to 1; 1≤b≤a, b is a positive integer; n is a positive integer greater than or equal to 1; 1≤i≤n, i is a positive integer; k is a positive integer greater than or equal to 1.

[0011] Optionally, the evaluation scores are divided into different ranges, and each range corresponds to a different evaluation result, and the evaluation results include comfortable, good, average, and poor.

[0012] Optionally, the typical operation scenarios include double-bottom welding of ships, climbing manhole covers, working at heights, and installation of pipelines in ship cabins.

[0013] Optionally, the key features include the size of the working area, the working posture of the construction workers, and the weight of the working tools.

[0014] Optionally, the multiple dimensions include operational accessibility, visibility, comfort, and safety.

[0015] The present invention also provides a human factors engineering evaluation method, comprising the following steps:

[0016] The human factors engineering evaluation training model is constructed by using the above-mentioned method for constructing the human factors engineering evaluation training model, wherein a confidence threshold is set in the human factors engineering evaluation training model;

[0017] Determine a work scenario to be evaluated, obtain a feature data set of the work scenario, and input the feature data set into the human factors engineering evaluation training model;

[0018] The human factors engineering evaluation training model evaluates the human factors engineering of the operating scenario and outputs the human factors engineering evaluation result in the operating scenario.

[0019] Optionally, it also includes:

[0020] The human factors engineering assessment training model outputs the human factors assessment result of the operation scenario and the corresponding confidence level;

[0021] The confidence level is compared with the confidence threshold. When the confidence level is greater than or equal to the confidence threshold, the human factors engineering assessment result is output; when the confidence level is less than the confidence threshold, the operation scenario is used as a new typical operation scenario, and the human factors assessment training model is reconstructed.

[0022] The present invention provides a human factors engineering training model, comprising:

[0023] A task receiving module, used to receive characteristic data of the human factors engineering operation scene to be evaluated;

[0024] An evaluation module, used for evaluating the received operation scenario;

[0025] A user interaction module, used to display the evaluation results output by the human factors evaluation training model;

[0026] A data storage module is used to store the operation scene information and the evaluation data.

[0027] Optionally, a comparison module is also included, which is used to compare the confidence level with a confidence threshold.

[0028] Optionally, the evaluation module includes: an accessibility evaluation module, a visibility evaluation module, a comfort evaluation module and a safety evaluation module.

[0029] As described above, the human factors engineering assessment method provided by the present invention has at least the following beneficial technical effects:

[0030] Improve the evaluation efficiency of ergonomics: The present invention applies artificial intelligence technology, especially deep learning and machine learning methods, to determine the characteristic data set of the work scene, input it into the ergonomics evaluation training model, and output the predicted evaluation results and confidence levels. Compared with traditional manual evaluation methods, the AI ​​model can quickly process a large number of complex work scenes without the need for complex simulation processes. Through automation and intelligent means, it reduces the complexity of manual operations and greatly shortens the evaluation time.

[0031] Improve assessment accuracy: This invention arranges construction workers of different age groups and skill levels to perform actual scoring in a real environment, constructs an evaluation data set based on multiple dimensions (operation accessibility, visibility, comfort, safety, etc.), and sets weights based on expert experience to make the assessment more in line with actual operating conditions.

[0032] The model has adaptive capabilities: The human factors engineering assessment training model can optimize itself through real-time data and feedback. When a new operating scenario appears or the confidence level of the assessment result is lower than the threshold, the model will be retrained with the new operating scenario feature data, and the model will automatically update and retrain. This ensures the accuracy and credibility of the assessment results, and has strong flexibility and intelligence. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 The diagram shows a method for constructing a human factors engineering assessment model provided in the first embodiment of the present invention.

[0034] Figure 2The display is the human factors engineering evaluation method provided by the second embodiment of the present invention.

[0035] Figure 3 Shown is a schematic diagram of the composition of the human factors engineering assessment model provided in the third embodiment of the present invention.

[0036] Reference numerals

[0037] 10. Task receiving module; 20. Evaluation module; 21. Accessibility evaluation module; 22. Visibility evaluation module; 23. Comfort evaluation module; 24. Safety evaluation module; 30. User interaction module; 31. Input end; 32. Output end; 33. Display end; 40. Data storage module; 50. Comparison module. DETAILED DESCRIPTION

[0038] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0039] It should be noted that the illustrations provided in this embodiment only illustrate the basic concept of the present invention in a schematic manner. Although the illustrations only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation, the form, quantity, positional relationship and proportion of each component in actual implementation can be changed at will under the premise of realizing the technical solution of this party, and the component layout form may also be more complicated.

[0040] Embodiment 1

[0041] This embodiment provides a method for constructing a human factors engineering assessment training model. Figure 1 As shown, the following steps are included to build a human factors engineering assessment training model:

[0042] S11: Construct a typical operation scenario set Pa;

[0043] S12: Select any typical operation scene Pb and extract key features tb of the typical operation scene i , construct feature data set Tb = {tb1, tb2...tb n};

[0044] S13: Arrange construction personnel to perform actual operations in the real environment of the typical operation scene, and perform multi-dimensional scoring on the operation scene Pb according to the feature data set to construct an initial score set Sb = {sb1, sb2...sb k};

[0045] S14: Obtain evaluation scores: Construct a weight set Wb = {wb1, wb2...wb k}, weighted scoring to obtain the evaluation score Ab = sb1*wb1+sb2*wb2…sb k *wb k ;

[0046] S15: Summarize the feature data set Tb to form a feature data set combination TT and an evaluation data set combination AA corresponding to the feature data set combination TT, TT = {T1, T2 ... Ta}, AA = {A1, A2 ... Aa};

[0047] Among them, a is a positive integer greater than or equal to 1; 1≤b≤a, b is a positive integer; n is a positive integer greater than or equal to 1; 1≤i≤n, i is a positive integer; k is a positive integer greater than or equal to 1.

[0048] Generally, in the construction of human factors engineering assessment training model, the following methods are mainly included: based on multi-classification problem machine learning algorithms, such as decision tree algorithm, random forest algorithm, support vector machine algorithm; based on deep neural network methods, such as multi-layer perceptron, convolutional neural network, recurrent neural network and other algorithms. Specifically, the present invention can construct a human factors engineering assessment training model based on multi-classification problem machine learning algorithms or using deep neural network methods.

[0049] S11: Construct a typical operation scenario set Pa; where a is a positive integer greater than or equal to 1.

[0050] Specifically, this embodiment takes the ship manufacturing scenario as an example and lists a typical operation scenario set Pa. Specifically, Pa includes: P1 ship double bottom welding; P2 manhole cover climbing; P3 high-altitude operation; P4 ship cabin pipeline installation; P5 small cabin equipment maintenance; P6 ship main engine installation and commissioning; P7 ship painting operation; P8 ship electrical system wiring and installation; P9 ship section assembling, P10 ship propeller installation and disassembly... operations involving small and confined spaces.

[0051] S12: Select any typical operation scene Pb and extract key features tb of the typical operation scene i , construct feature data set Tb = {tb1, tb2...tb n}; wherein, 1≤b≤a, b is a positive integer; n is a positive integer greater than or equal to 1; 1≤i≤n, i is a positive integer.

[0052] Generally, the key feature tb iThese factors include: the size of the operating area (including the length, width and height of the operating area), the environment of the operating area (including the temperature, humidity, noise, vibration and lighting conditions of the operating area), the weight of the operating tools, the working posture of the construction workers, the professional skills and experience differences of the construction workers, the rationality of the operation plan and scheduling, and other factors that affect ship operations.

[0053] Specifically, taking the double bottom welding operation of P1 ship (b=1) as an example, a feature data set T1={t11, t12...t1 n Specifically, its key features, T11 include: t11 the length, width and height of the working area, t12 the posture of the construction workers during the welding process, t13 the closest distance from the welding part to the construction workers in their construction posture, t14 the weight of the welding gun and other tools, etc. ...}.

[0054] Specifically, taking the P2 manhole cover climbing operation (b=2) as an example, its key features constitute the feature data set T2={t21, t22...t2 n Specifically, T2 = {t21 climbing height, t22 diameter of manhole cover, t23 complexity of surrounding environment (such as whether there are obstacles, light conditions, etc.), t24 type and location of climbing assistance device...}.

[0055] Specifically, taking P3 high-altitude work (b=3) as an example, its key features constitute the feature data set T3={t31, t32...t3 n Specifically, T3 = {t31 working height, t32 stability of the working platform, t33 the configuration of safety protection facilities, t34 the lifting method and difficulty of tools and materials, t35 the impact of natural environmental factors such as wind on the work, etc.}.

[0056] Specifically, taking the P4 ship cabin interior pipe installation (b=4) as an example, its key features constitute the feature data set T4={t41, t42...t4 n Specifically, T4 = {t41 cabin space size and shape, t42 pipeline diameter and length, t43 pipeline layout complexity, t44 installation location relative to other equipment or structures, t45 accessibility of connection parts...}.

[0057] Specifically, taking P5 small cabin equipment maintenance (b=5) as an example, its key features constitute the feature data set T5={t51, t52...t5 n Specifically, T5 = {t51 the narrowness of the cabin (length, width, height), t52 the type and size of the equipment, t53 the type of tools required for maintenance and the operating space requirements, t54 the accessibility of the equipment failure site, t55 the complexity of the maintenance operation...}.

[0058] Specifically, the acquired feature data set Tb is preliminarily screened. For example, if the data recorded by the temperature sensor far exceeds the normal welding temperature range, it may be a sensor failure or abnormal situation), or some missing data points are also processed at this stage, or some obviously wrong data, such as the work scene layout information obtained by image recognition technology, which obviously does not match the actual data (such as the size and position of the identified object do not conform to physical common sense), can be excluded at this stage to ensure that the data subsequently input into the evaluation training model is accurate and reliable.

[0059] S13: Arrange construction personnel to perform actual operations in the real environment of the typical work scene, and perform multi-dimensional scoring on the work scene Pb according to the feature data set to construct an initial score set Sb = {sb1, sb2...sbk}; k is a positive integer greater than or equal to 1.

[0060] Generally, construction workers should cover all age groups as much as possible, including different levels such as skilled workers and novices. Construction workers give the work score of the work scene according to the actual work actions and work time under real circumstances. Among them, the score can be considered from the dimensions of operation accessibility, operation visibility, operation comfort, and operation safety. Construction workers can wear sensing devices to assist in the evaluation of comfort, etc., and evaluate and score based on the subjective feelings of the workers.

[0061] Specifically, since the key features of each task are not exactly the same, the key feature tb i There are various forms of evaluation and scoring. Generally, the evaluation and scoring methods include quantitative scoring and qualitative scoring and other methods. Quantitative scoring methods include: direct numerical scoring: for quantifiable characteristics such as the size of the operating area and the weight of the tool, for example, the operating height of the ship's cabin is divided into several intervals, 1-3 meters is 4 points, 3-5 meters is 3 points, 5-7 meters is 2 points, and 7 meters and above is 1 point. For the weight of the tool, 5 kilograms is a gradient, and every increase of 5 kilograms is reduced by 1 point, starting from 10 points, clearly quantifying the operating load brought by the weight of the tool; it also includes proportional scoring: if the complexity is 1.5 times the standard layout, it is 3 points, 2 times is 2 points, and 3 times and above is 1 point. The score is determined by the relative proportional relationship, highlighting the impact of the differences in pipeline layout in different scenarios on the operation. Qualitative scoring methods include descriptive grade scoring: for example, in the visibility evaluation, it can be divided into four levels of "clearly visible, basically visible, partially visible, and unclear", corresponding to 4-1 points respectively. Construction workers make subjective judgments based on actual observations and select the corresponding level, converting the blurred visibility conditions into quantifiable scores.

[0062] Specifically, in this embodiment, the descriptive evaluation method in the qualitative evaluation method is adopted, and the scoring criteria are refined as follows: the operational accessibility is divided into four levels: easily accessible, relatively easy to reach, barely accessible, and difficult to reach, corresponding to 4-1 points respectively; visibility is divided into four levels: clearly visible, basically visible, partially visible, and difficult to see, corresponding to 4-1 points; comfort is divided into four levels: comfortable, relatively comfortable, general, and uncomfortable according to the degree of physical fatigue and discomfort of the construction personnel, corresponding to 4-1 points; safety is divided into four levels: safe, relatively safe, risky, and dangerous according to the size of potential risks and the possibility of accidents, corresponding to 4-1 points. Through such detailed scoring criteria, the key features of the operation scene can be more accurately quantitatively evaluated, providing a more reliable data basis for the subsequent construction of the evaluation model. Therefore, for different operation tasks, the evaluation scores of the key features by the construction personnel are not exactly the same.

[0063] Specifically, taking the double bottom welding operation of P1 ship (b=1) as an example, the construction personnel are arranged to perform welding operations in the real double bottom space. According to the feature data set, the operation scene Pb is scored in multiple dimensions, including: for operational accessibility, if the length, width and height of the key feature t11 operation area are small, the space is relatively cramped, and the construction personnel are affected in performing large-scale movements, then the accessibility score is low (such as S11 = 2 points); for operational visibility, if the key feature t13 is far from the closest distance between the welding part and the construction personnel in the construction posture (or other key features: insufficient light and complex spatial layout make it difficult to observe the welding part), then the visibility score is low (such as S12 = 2 points); for operational comfort, if the key feature t12 construction personnel maintain an uncomfortable or fatigue-prone posture for a long time during the welding process, then the comfort score will be low (such as S13 = 1 point); for operational safety, if the welding quality and safety are affected by unstable posture, then the safety score is low (such as S14 = 2 points); that is, for the initial score set S1 = {s11, s12, s13, s14} = {2, 2, 1, 2} for the double bottom welding operation of the ship P1.

[0064] Specifically, taking the P2 manhole cover climbing operation (b=2) as an example, the scores of the key features are: if t21 the climbing height is high and t22 the diameter of the manhole cover is small, the accessibility score is low (such as S21=2 points) and the comfort score is low (such as S23=1 point); t23 the complexity of the surrounding environment is high and the visibility score is low (such as S22=3 points); if there is a slippery situation, the scores in terms of stability and safety will be significantly reduced, and the safety score is low (such as S24=2 points); that is, for the initial score set S2={s21, s22, s23, s24}=(2,3,1,2) for the P2 manhole cover climbing operation.

[0065] Specifically, taking P3 aerial work (b=3) as an example, the scores of key features are: t31 the working height is relatively high, t32 the stability of the working platform is poor, t34 the lifting method of tools and materials is difficult, and if the working platform shakes significantly, it will greatly increase the difficulty of operation, so the accessibility score is low (such as S1=2 points); t33 the safety protection facilities are not adequately equipped, so the safety score is low (such as S4=1 point); the visibility is high, so the visibility score is high (such as S2=4 points); the comfort is high (such as S3=4 points); S3={s31, s32, s33, s34}=(2,4,4,1).

[0066] Specifically, taking the piping installation in the P4 ship cabin (b=4) as an example, if the diameter of the t42 pipeline is narrow and the length is long, the accessibility score is low (such as S1=1 point); the comfort score is low (such as S3=1 point); the t41 cabin space is narrow, the t43 pipeline layout is complex, and the operating space of the t45 connection part is small, which will increase the difficulty of installation, then the visibility score is low (such as S2=2 points); the safety is high (such as S4=3); that is, S4={s41, s42, s43, s44}=(1,2,1,3).

[0067] Specifically, taking the P5 small cabin equipment maintenance (b=5) as an example, if the accessibility score is low (such as S1=1 point), the t52 equipment visibility is also low (such as S2=2 points); the t51 cabin is too small, which restricts the use of maintenance tools and personnel movements, and the comfort score is low (such as S3=1 point); the safety is high (such as S4=3); that is, S5={s51, s52, s53, s54}=(1,2,1,3).

[0068] S14: Obtain evaluation score A: Set weight set Wb = {wb1, wb2...wb k}, weighted scoring to obtain the evaluation score Ab = sb1*wb1+sb2*wb2…sb k *wb k ;

[0069] First, the evaluation weight set W needs to be set. The evaluation weight set is usually set by experts based on rich professional knowledge and practical experience, and is set by comprehensively considering the importance of multiple dimensions in actual operations. Therefore, the evaluation weight set W for each task is not exactly the same. Specifically, in this embodiment, the weight set Wb = {wb1, wb2, wb3, wb4}; wb1 is the accessibility weight; wb1 is the visibility weight; wb1 is the comfort weight; wb1 is the safety weight. Weighted scoring evaluation score Ab = sb1*wb1+sb2*wb2+sb3*wb3+sb4*wb4.

[0070] Specifically, in the double bottom welding operation of P1 ship (b = 1), if welding safety is considered to be of vital importance, w14 = 0.4 can be set; welding accessibility is also relatively critical, w11 = 0.3; welding visibility and comfort are relatively inferior, w12 = 0.15 and w13 = 0.15 respectively; that is, W1 = (0.3, 0.15, 0.15, 0.4). The weighted scoring evaluation score for the double bottom welding operation of the ship is: A1 = s11*w11+s12*w12+s13*w13+s14*w1 4＝ 2*0.3+1*0.15+1*0.15+2*0.4=1.7.

[0071] Specifically, taking the P2 manhole cover climbing operation (b=2) as an example, if the climbing safety weight w24=0.5; climbing accessibility weight w21=0.3; surrounding environment visibility weight w22=0.2; comfort weight w23=0; that is, W2=(0.3, 0.2, 0, 0.5). Then for the manhole cover climbing operation, the weighted evaluation score is A2=s21*w21+s22*w22+s23*w23+s24*w2 4＝ =0.3*2+0.2*2+0*1+0.5*2=2.

[0072] Specifically, taking P3 aerial work (b=3) as an example, if the safety weight w34=0.45, the accessibility weight w31=0.25, the visibility weight w32=0.1, and the comfort weight w33=0.2, that is, W3=(0.25, 0.1, 0.2, 0.45), the weighted scoring evaluation score for aerial work is: A3=s31*w31+s32*w32+s33*w33+s34*w34=0.25*2+0.1*3+0.2*4+0.45*1=1.9.

[0073] Specifically, taking the installation of pipelines in P4 ship cabins (b=4) as an example, if the accessibility weight w41=0.35, the visibility weight w42=0.2, the comfort weight w43=0.1, and the safety weight w44=0.35, that is, W4=(0.35, 0.2, 0.1, 0.35), the weighted scoring evaluation score A4=s41*w41+s42*w42+s43*w43+s44*w44=0.35*1+0.2*1+0.1*1+0.35*3=1.6.

[0074] Specifically, taking P5 small cabin equipment maintenance (b=5) as an example, if the accessibility weight w51=0.4, the visibility weight w52=0.2, the comfort weight w53=0.1, and the safety weight w54=0.3; that is, W5=(0.4, 0.2, 0.1, 0.3), then the weighted scoring evaluation score A=s51*w51+s52*w52+s53*w53+s54*w54=0.4*1+0.2*1+0.1*1+0.3*3=1.6.

[0075] Optionally, according to the evaluation score Ab, the evaluation score is divided into different ranges, and the evaluation results of each range are corresponding. Specifically, the evaluation results include comfortable, good, general, and poor. Specifically, if Ab>3, the output evaluation result is comfortable; if 2<Ab≤3, the output evaluation result is good; if 1<Ab≤2, the output evaluation result is general; Ab≤1, the output result is poor. Specifically, in this embodiment, for the P1 ship double bottom welding operation (b=1), A1 is between 1-2, so the output result of the ship double bottom welding operation provided in this embodiment is general; for the P2 manhole cover climbing operation (b=2), between 1<A2≤2, the output evaluation result is general; for the P3 high-altitude operation (b=3), between 1<A3≤2, the output evaluation result is general. For the P4 ship cabin interior pipeline installation (b=4), between 1<A4≤2, the output evaluation result is general; for the P5 small cabin equipment maintenance (b=5), between 1<A≤2, the output evaluation result is general.

[0076] S15: Aggregate the feature data set Tb to form a feature data set combination TT and an evaluation data set combination AA corresponding to the feature data set combination TT, where TT = {T1, T2...Ta}, AA = {A1, A2...Aa}.

[0077] Specifically, the feature data collected for each operation scene are integrated to form a feature data set combination TT. For example, for a ship construction project, there are 100 double-bottom welding operation samples, 80 manhole cover climbing operation samples, 120 high-altitude operation samples, 90 ship cabin pipe installation samples, and 70 small cabin equipment maintenance samples. The feature data of these samples are sorted into feature data sets such as T1, T2, T3, T4, and T5, and then combined into a complete feature data set combination TT.

[0078] Specifically, after construction workers perform practical operations in a real environment and score them, they are summarized to form an evaluation data set combination AA. For example, for the above 100 double-bottom welding operation samples, 100 evaluation scores are obtained, and the method of taking the average of multiple scores is adopted to remove outliers. A1 is obtained by comprehensive calculation through a specific mathematical model, such as the weighted average method, considering the importance of different scoring dimensions. By analogy, for 80 manhole cover climbing operation samples, 120 high-altitude operation samples, 90 ship cabin pipeline installation samples and 70 small cabin equipment maintenance samples, the same scoring, error processing and comprehensive calculation are performed to obtain A2, A3, A4, and A5. Finally, A1, A2, A3, A4, and A5 are summarized into the evaluation data set combination AA.

[0079] In addition, under normal circumstances, different shipbuilding companies have different requirements for human factors engineering assessment results. The confidence threshold can be adjusted according to the business needs and tolerance of the company. For example, for operating scenarios with extremely high safety requirements, the company may choose to set a higher confidence threshold to ensure the accuracy of the prediction results.

[0080] Generally, methods for determining confidence thresholds include the mean + standard deviation method based on data distribution, business experience and expert judgment, and cross-validation. Some evaluation methods can be used to understand the performance of the model at different confidence levels. Specifically, in actual operations, the determination of the confidence threshold requires multiple iterations and adjustments.

[0081] The method for constructing the human factors engineering evaluation training model provided in this embodiment forms a complete evaluation data set A by aggregating similar calculations in a large number of different operation scenarios, and comprehensively considers the impact of different factors on human factors engineering from multiple dimensions, so as to more accurately evaluate the human factors engineering status in the shipbuilding process and ensure the safety of operators and the efficient operation.

[0082] Embodiment 2

[0083] This embodiment provides a human factors engineering evaluation method for evaluating human factors engineering in typical shipbuilding scenarios, such as Figure 1 As shown, the method comprises the following steps:

[0084] S1: Adopting the method for constructing a human factors engineering evaluation training model described in Example 1, a human factors engineering evaluation training model is constructed, wherein a confidence threshold is set in the human factors engineering evaluation training model.

[0085] The method for constructing a human factors engineering assessment training model includes the method for constructing a human factors engineering assessment training model described in Example 1. Generally, the confidence threshold in the model is not fixed, but is set according to the actual operating scenario. Taking the double-bottom welding of ships as an example, when setting the confidence threshold, it should be considered that: double-bottom welding involves complex spatial layout, high-temperature operations and harmful gas emissions, and its confidence threshold setting needs to take into account the changes in multiple factors such as welding accessibility, visibility, comfort and safety under different operating conditions.

[0086] S2: Determine a work scenario that needs to be evaluated, obtain a feature data set of the work scenario, and input the feature data set into the ergonomics evaluation training model.

[0087] Specifically, the tools for obtaining feature data sets include: sensors: position and displacement sensors, mechanical sensors, environmental sensors, posture sensors; image acquisition and recognition, high-definition cameras, image recognition software; manual recording and measurement, traditional tape measures, calipers, levels, etc., inspection sheets and record sheets, etc., combined with manual records, to comprehensively obtain the precise dimensions of the work area, detailed parameters of the work tools, real-time posture data of the construction personnel, etc.

[0088] For example, in the double-bottom welding operation scene of a ship, high-precision laser ranging sensors and three-dimensional modeling technology are used to measure the length, width and height of the operation area, which are 5 meters, 3 meters and 1.5 meters respectively. The weight of the welding gun is obtained by the pressure sensor as 3 kg. With the help of wearable motion capture equipment, the body posture of the construction workers is monitored in real time, and the posture data such as the bending angle of the arm during welding is between 30°-120° and the inclination angle of the body is maintained at about 10°-20°. At the same time, the environmental information such as the temperature of the welding environment is 35°C and the humidity is 70% is manually recorded. Therefore, the feature data set T = {operation area length 5 meters, operation area width 3 meters, operation area height 1.5 meters, welding gun weight 3 kg, construction worker arm bending angle range 30°-120°, body inclination angle 10°-20°, welding environment temperature 35°C, welding environment humidity 0%}. The feature data set T is input into the human factors engineering evaluation training model to evaluate the human factors engineering of the double-bottom welding operation scene of the ship.

[0089] S3: The human factors engineering evaluation training model evaluates the human factors engineering of the operating scenario and outputs the human factors engineering evaluation result in the operating scenario.

[0090] Specifically, the human factors engineering assessment training model analyzes and evaluates the input operation scene feature data set T based on the built-in multi-classification problem machine learning algorithm or deep neural network architecture. Thus, the human factors engineering assessment results and the corresponding confidence values ​​are obtained. Taking the ship cabin pipeline installation operation as an example, the model will comprehensively consider factors such as the size of the cabin space, pipeline layout, accessibility and comfort when operating tools, and give evaluation results such as "good", "average" or "poor".

[0091] S4: The human factors engineering assessment training model outputs the human factors assessment result of the operation scene and the corresponding confidence level at the same time. The confidence level is compared with the confidence level threshold. If the confidence level is higher than the confidence level threshold, the human factors engineering assessment result is output; if the confidence level is lower than the confidence level threshold, the operation scene is used as new feature data to reconstruct the human factors assessment training model.

[0092] Comparing the confidence level with the confidence threshold is a key step in determining whether the evaluation result is reliable. When the confidence level is higher than the confidence threshold, it indicates that the model's evaluation results for the operation scenario have high reliability and credibility, and the human factors engineering evaluation results can be directly output at this time. For example, in the ship manhole cover climbing operation scenario, if the evaluation result output by the model is "good" and the confidence level reaches 0.8 (assuming the threshold is 0.7), it can be judged based on this result that the operation scenario basically meets the requirements in terms of human factors engineering, and the current operation method and safety measures can be maintained. At the same time, the relevant data is recorded and analyzed to provide a reference for subsequent similar operations.

[0093] If the confidence is lower than the confidence threshold, it indicates that the human factors engineering evaluation model has a large uncertainty in its evaluation of the work scene. At this time, the work scene needs to be treated as a new work scene, its feature data set needs to be obtained, and the human factors engineering evaluation model needs to be trained to continuously expand and improve the model. Specifically, when a new work scene appears, the feature data set of the new work scene is also obtained, and the human factors engineering evaluation model is then trained.

[0094] For example, a new type of automated welding equipment was introduced in the shipbuilding process. Its operation scenario is significantly different from the traditional welding method, resulting in a low confidence level in the initial model's assessment. By reintegrating the feature data of the new operation scenario into the training set and retraining the model, the model can learn new operation modes and influencing factors, thereby improving the accuracy and reliability of the assessment of similar new scenarios and promoting the continuous development and innovation of human factors engineering assessment technology in shipbuilding and other related fields.

[0095] In summary, this embodiment provides a human factors engineering assessment method, which can output assessment results and confidence levels by inputting a feature data set of an operation scenario, thereby greatly reducing the complexity of manual operations, improving the efficiency of human factors engineering assessment, and providing strong support for human factors engineering assessment in the shipbuilding process.

[0096] Embodiment 3

[0097] This embodiment also provides a human factors engineering assessment training model, such as Figure 3 As shown, the human factors engineering training model provided in this embodiment includes: a task receiving module 10, an evaluation module 20, a user interaction module 30, a data storage module 40 and a comparison module 50. The task receiving module 10 is used to receive the feature data of the human factors engineering operation scene to be evaluated; the evaluation module 20 is used to evaluate the received operation scene; the user interaction module 30 is used to display the evaluation result output by the human factors evaluation training model; the data storage module 40 is used to store the operation scene information and the evaluation information, including the original feature data, the intermediate results in the evaluation process and the final evaluation report; the comparison module 50 is used to compare the confidence level with the confidence threshold.

[0098] Specifically, the task receiving module 10 receives characteristic data of the ergonomic work scene to be evaluated, including but not limited to data sent in real time by various sensors installed at the shipbuilding site (such as temperature sensors, pressure sensors, position sensors, etc.), data manually input by construction personnel (such as environmental descriptions of the work site, feedback on operating difficulties, etc.), and data transmitted from other systems (such as the production management system and equipment management system of the shipbuilding enterprise); secondly, the task receiving module 10 is also provided with a task parsing module 11, which is used to convert data in different formats into a unified format acceptable to the evaluation module 20; in addition, the task receiving module 10 also includes a data screening module 12, which can perform preliminary classification and inspection of the characteristic data to ensure the accuracy of the data and mark the values ​​that are obviously beyond the normal range.

[0099] Specifically, the evaluation module 20 includes an accessibility evaluation module 21, a visibility evaluation module 22, a comfort evaluation module 23, and a safety evaluation module 24. The accessibility evaluation module 21 evaluates the accessibility of personnel to various equipment, controls, or work areas in the work scene, including evaluating whether the operator can easily and quickly reach the work place, and whether there are any obstructing factors (such as obstacles, narrow spaces, etc.); the visibility evaluation module 22 evaluates the visibility of the work environment, equipment, and important information (such as signs, display screens, warning signs, etc.) of the staff in the work scene, including evaluating the clarity of the environment, the rationality of the equipment, and whether there are problems such as line of sight obstruction; the comfort evaluation module 23 evaluates the comfort of the operators in the work scene, including the impact of factors such as body posture, workload, and ambient temperature and humidity on the personnel; the safety evaluation module 24 evaluates the safety hazards and risk points in the work scene, including injuries or accidents that may be caused by equipment failure, operating errors, environmental factors, etc.

[0100] Specifically, the user interaction module 30 includes an input terminal 31, an output terminal 32 and a display terminal 33. The input terminal 31 can input data to the human factors engineering assessment training model, set the confidence threshold, and adjust parameters. Specifically, parameter adjustment includes: algorithm selection, weight distribution, etc. The output terminal 32 is used to output the assessment results, output feedback prompts (such as problems found during the assessment process, recommended improvement measures, etc.), and can also export data: users can export the assessment results to files such as Excel tables, PDF reports, etc. through the output terminal for further analysis and application. The display terminal 33 is provided with a visual interface so that users can intuitively understand the assessment results. The display terminal also supports user interactive operations, such as zooming in / out charts, switching display modes, etc., so that users can have a deeper understanding of the assessment results.

[0101] Specifically, the comparison module 50 is used to compare the confidence level with the confidence threshold. If the confidence level calculated by the ergonomics evaluation training model is less than the confidence threshold, the output result is unreliable, and the characteristic information of the operation scene needs to be used as a new task to rebuild the ergonomics training model; if the confidence level is greater than or equal to the confidence threshold, the output result has a higher degree of certainty.

[0102] The above embodiments are merely illustrative of the principles and effects of the present invention, and are not intended to limit the present invention. Anyone familiar with the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by a person of ordinary skill in the art without departing from the spirit and technical concept disclosed by the present invention shall still be covered by the claims of the present invention.

Claims

1. A method for constructing a human factors engineering assessment training model, characterized in that: The following steps are involved: Construct a typical operation scenario set Pa; Select any typical operation scene Pb and extract the key features tb of the typical operation scene i , construct feature data set Tb = {tb1, tb2...tb n }; Arrange construction personnel to perform actual operations in the real environment of the typical operation scene, and perform multi-dimensional scoring on the operation scene Pb according to the feature data set to construct an initial score set Sb = {sb1, sb2...sb k }; Get evaluation scores: Construct weight set Wb = {wb1, wb2...wb k }, weighted scoring to obtain evaluation scores Ab=sb1*wb1+sb2*wb2……sb k *wb k ; Aggregate the feature data set Tb to form a feature data set combination TT and an evaluation data set combination AA corresponding to the feature data set combination TT, TT = {T1, T2 ... Ta}, AA = {A1, A2 ... Aa}; Wherein, a is a positive integer greater than or equal to 1; 1≤b≤a, b is a positive integer; n is a positive integer greater than or equal to 1; 1≤i≤n, i is a positive integer; k is a positive integer greater than or equal to 1.

2. The method for constructing a human factors engineering assessment training model according to claim 1, characterized in that: The evaluation scores are divided into different ranges, and each range corresponds to a different evaluation result, and the evaluation results include comfortable, good, average, and poor.

3. The method for constructing a human factors engineering assessment training model according to claim 1, characterized in that: The typical working scenarios include double-bottom welding of ships, climbing manhole covers, working at heights, and installing pipes in ship cabins.

4. The method for constructing a human factors engineering assessment training model according to claim 1, characterized in that: The key features include the size of the work area, the working posture of the construction personnel, and the weight of the work tools.

5. The method for constructing a human factors engineering assessment training model according to claim 1, characterized in that: The multiple dimensions include operational accessibility, visibility, comfort, and safety.

6. A human factors engineering evaluation method is used to evaluate the human factors engineering of typical shipbuilding operation scenarios, characterized in that: The following steps are involved: Adopting the method for constructing a human factors engineering assessment training model as described in any one of claims 1 to 5, a human factors engineering assessment training model is constructed, wherein a confidence threshold is set in the human factors engineering assessment training model; Determine an operation scenario to be evaluated, and obtain a feature data set of the operation scenario; Inputting the feature data set into the human factors engineering assessment training model; The human factors engineering evaluation training model evaluates the human factors engineering of the operating scenario and outputs the human factors engineering evaluation result of the operating scenario.

7. The human factors engineering evaluation method according to claim 6, characterized in that: Also includes: The human factors engineering assessment training model outputs the human factors assessment result of the operation scenario and the corresponding confidence level; The confidence level is compared with the confidence threshold. When the confidence level is greater than or equal to the confidence threshold, the human factors engineering assessment result is output; when the confidence level is less than the confidence threshold, the operation scenario is used as a new typical operation scenario, and the human factors assessment training model is reconstructed.

8. A human factors engineering assessment training model for conducting human factors engineering assessment on shipbuilding scenarios, comprising: A task receiving module, used to receive characteristic data of the human factors engineering operation scene to be evaluated; An evaluation module, used for evaluating the received operation scenario; A user interaction module, used to display the evaluation results output by the human factors evaluation training model; A data storage module is used to store the operation scene information and the evaluation data.

9. The human factors engineering assessment training model according to claim 8, characterized in that: Also includes: The comparison module is used to compare the confidence level with the confidence threshold.

10. The human factors engineering evaluation method according to claim 8, characterized in that: The evaluation modules include: an accessibility evaluation module, a visibility evaluation module, a comfort evaluation module and a safety evaluation module.

Images