Collaborative quality control method for preventing subjective human factor failure in nuclear power plant construction

Abstract

The application discloses a kind of collaborative quality control methods for preventing nuclear power construction subjective human factor failure, subjective human factor failure process model and safety control structure are constructed based on STAMP method to collaborative quality witnessing process, the formation path of the subjective human factor failure and safety control behavior of each participant in collaborative quality witnessing process is analyzed, and multiple risk factors of the subjective human factor failure are identified according to its formation path;The importance degree of each risk factor in preventing subjective human factor failure is sorted by using DEMATEL-ISM method, and the correlation logic between each risk factor is analyzed by constructing multi-level hierarchical model of risk factors;The technical facility factor in risk factor is selected for optimization and improvement, and the collaborative quality witnessing process and system architecture for preventing subjective human factor failure are designed.The application proposes a kind of collaborative quality intelligent witnessing method for preventing subjective human factor failure, which provides decision support for site construction management and safety quality supervision.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of nuclear power construction collaborative quality, in particular to a collaborative quality control method for preventing subjective human factor failure in nuclear power construction. BACKGROUND

[0002] The active and orderly development of nuclear power is in line with the energy system construction of the double carbon target, but because of the risk of radioactive material leakage, it poses a safety threat to all participating enterprises, the local society and the ecological environment, therefore the construction quality and safe operation of nuclear power equipment are the fundamental guarantee for enterprise benefits and public safety. Nuclear power equipment is a large and complex product with complex structure, high engineering and technical content, and high integration of parts, in addition to complying with the ISO9001 quality standard, it must also comply with the extremely strict nuclear safety quality law HAF003. In order to ensure that the items and services outsourced to following and participating enterprises meet the requirements of the nuclear safety quality law and follow the nuclear safety quality culture, nuclear power construction enterprises use multi-level supervision collaborative quality witnessing activities to complete the quality verification of outsourced items and services of nuclear power equipment in practice.

[0003] The collaborative quality witnessing activities in the nuclear power construction phase involve thousands of enterprises participating and hundreds of thousands of parts, and have characteristics such as dispersed process nodes, numerous participants, large quantities of materials, and multi-source heterogeneous information. In order to achieve the quality target, the provisions on quality witnessing in the quality assurance outline are strict and have complicated items, resulting in a complex quality witnessing interaction process, high work intensity and high collaboration difficulty, and subjective human factor failure is prone to occur. Moreover, human factor failure is also one of the main factors of nuclear power accidents, and has always been the focus of quality control. Therefore, studying the formation mechanism of subjective human factor failure of quality witnessing participants and preventing potential quality and safety risks caused by it are urgent problems to be solved in nuclear power construction quality witnessing. SUMMARY

[0004] In view of the above problems of the prior art, the technical problem to be solved by the present application is how to provide a collaborative quality control method for preventing subjective human factor failure in nuclear power construction to reduce potential quality and safety risks in the nuclear power construction process.

[0005] In order to solve the above technical problems, the technical scheme adopted by the present application is as follows:

[0006] A collaborative quality control method for preventing subjective human factor failure in nuclear power construction, comprising the following steps:

[0007] (1) Constructing a subjective human factor failure process model and a safety control structure based on the STAMP method for the collaborative quality witnessing process, analyzing the formation path of subjective human factor failure and safety control behavior of each participant in the collaborative quality witnessing process, and identifying multiple risk factors of subjective human factor failure according to the formation path of subjective human factor failure and safety control behavior;

[0008] (2) The importance degree of each risk factor in preventing subjective human error is calculated by using the DEMATEL-ISM method, the correlation logic between the risk factors is analyzed by constructing a multi-level hierarchical model of the risk factors, and the risk factors are divided into organizational factors, team factors, personal factors and technical facility factors;

[0009] (3) The technical facility factors are selected for optimization and improvement, and a collaborative quality witnessing process and system architecture for preventing subjective human error are designed.

[0010] As optimization, in step (1), the subjective human error process model comprises a controller, an executor, a controlled object and a sensor, the executor receives instructions and feeds back information to make the controlled object in a controllable state, the controller comprises a general contractor, an owner and a supervision company, the executor comprises a witnessing supervised party, the controlled object comprises a component or a process, and the sensor comprises a quality inspection instrument.

[0011] The safety control structure comprises a management layer, an execution layer and a physical layer in order of priority, each layer controls by imposing safety constraints on a lower layer subsystem, and the lower layer subsystem performs relevant operations under the safety constraints of the upper layer subsystem and feeds back implementation information to the upper layer subsystem.

[0012] As optimization, in step (2), when calculating and analyzing the importance degree and correlation logic of the risk factors, the following steps are adopted:

[0013] (a1) The influence degree values between the factors are determined, the risk factors are scored by using the Delphi method, when i = j, there is α ij = 0, a direct influence matrix A' is obtained; the scores in the direct influence A' are replaced by triangular fuzzy numbers by using a five-granularity fuzzy evaluation language set, the triangular fuzzy numbers are de-fuzzified by using the barycenter method, a transformed direct influence matrix B' is obtained, the risk factors are scored multiple times by repeating the above method to obtain multiple transformed direct influence matrices B', and an average matrix B is obtained by averaging the multiple transformed direct influence matrices B';

[0014] (a2) The average matrix B is normalized to obtain a standard influence matrix C, as formula 1:

[0015]

[0016] The comprehensive influence matrix T is calculated, T = t ij , as formula 2:

[0017] T = C (I-C) -1 (2)

[0018] Determine the indirect influence relationship between each risk factor, and then calculate the influence degree r of the risk factor i As formula 3:

[0019]

[0020] The affected degree s i As formula 4:

[0021]

[0022] The center degree m i As formula 5:

[0023] m i =r i +s i (5)

[0024] The reason degree n i As formula 6:

[0025] n i =r i -s i (6)

[0026] To determine the importance and attributes of each risk factor;

[0027] (a3) Calculate the overall influence relationship matrix H, where I is the unit matrix, H = I + T, establish the reachable matrix K, given threshold λ, as formula 7:

[0028]

[0029] k ij =1 indicates that factor a i Can influence factor a j , k ij =0 indicates that factor a i Can't influence factor a j , threshold λ is added according to the mean α and standard deviation β of all elements in the comprehensive influence matrix T;

[0030] According to the reachable matrix K, the hierarchical division is carried out, first determine the reachable set P i And the precedence set Q i , the intersection of the two is U i , use the result priority extraction method to carry out hierarchical division, if P i =U i , the element is the first level, and the corresponding row and column in the reachable matrix K are deleted, and the above process is repeated until all risk factors are allocated, draw the risk factor multi-level hierarchical model, and analyze the correlation logic between each risk factor from the risk factor multi-level hierarchical model.

[0031] As optimization, in step (3), the collaborative quality witnessing process comprises the following steps:

[0032] (b1) The operator completes the collection of equipment, personnel and environment data through the handheld terminal device, and realizes real-time intelligent perception and interconnection of the witnessing process data information through the Internet of Things technology;

[0033] (b2) Based on the processing, storage, analysis and mining capacity of big data, the data of user qualification review, personnel identity authentication, quality photo, video record and test report of the process generated by the collaborative quality witnessing process are cleaned, dimensionally reduced and mined to identify the equipment quality state, witnessing execution progress and personnel operation behavior of the nuclear power construction process;

[0034] (b3) Establish a quality intelligent witnessing platform to complete the quality witnessing management, quality test management and system basic management functions of the quality intelligent witnessing platform, integrate the quality data of the witnessing process, and display and interact in the quality intelligent witnessing platform in a visual manner, and identify, supervise and control the subjective human factor failure risk in the quality witnessing process of the nuclear power equipment.

[0035] As optimization, in step (3), the architecture comprises an edge perception layer, a center processing layer and an application service layer;

[0036] The edge perception layer includes physical resource elements and data collection, the physical resource elements include nuclear power items, quality test processes and quality witnessing personnel; the data collection collects objective data and information of the physical resource element situation through sensing devices, quality test instruments and video monitoring, the objective data and information of the physical resource element situation include basic attributes and location information of the item, process requirements of the process, detection results of the quality, identity information and human-computer interaction behavior of the quality state personnel;

[0037] The center processing layer converts the information data used and interacted in the witnessing business in the nuclear power construction into standardized data through cleaning, integration and dimension reduction processing through big data technology, analyzes and mines through natural semantic processing and language pre-training model, completes data specification packaging, completes real-time transmission based on data standard protocol, forms end-to-end cross-enterprise information docking and finally integrates in a unified data space;

[0038] The application service layer provides intelligent services supporting the collaborative quality witnessing management business with process control as the goal, and the application service layer includes business process improvement, knowledge push optimization, execution personnel supervision and file security service;

[0039] The business process improvement is carried out around the electronic business process of collaborative quality witnessing, and promotes the standardization, systematization and intelligentization of collaborative quality witnessing;

[0040] Knowledge push optimization provides procedures and experience knowledge with collaborative quality witness in compliance with legal regulations and technical route requirements;

[0041] Performing personnel supervision is to confirm and supervise quality witness personnel;

[0042] File security service provides quality process files, archive file records and effectiveness verification with collaborative quality witness, and provides basis for traceability of quality problems.

[0043] Compared with the prior art, the present application has the following advantages: the present application constructs a subjective human factor failure process model and a safety control structure based on the STAMP method, analyzes safety control behaviors of each participant in the quality witness process and a formation path of subjective human factor failure events, and identifies potential risk factors of subjective human factor failure according to the same; the DEMATEL-ISM method is used to sort the risk influencing factors, and the influence of key risk factors on the subjective human factor failure of collaborative quality witness is revealed; a collaborative quality intelligent witness method for preventing subjective human factor failure is proposed, and an intelligent witness process and a technical system architecture are described so as to establish technical and management barriers; illegal or unsafe behaviors are identified in advance, and the occurrence of quality problems or potential hazards is eliminated, thereby playing the role of an intelligent barrier for preventing subjective human factor failure, and providing decision support for field construction management and safety quality supervision. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 A subjective human factor failure process model in the present application;

[0045] Figure 2 A subjective human factor failure prevention hierarchical control structure in the present application;

[0046] Figure 3 A risk factor multi-level hierarchical model in the present application;

[0047] Figure 4 A collaborative quality witness process of subjective human factor failure in the present application;

[0048] Figure 5 A system architecture of a cross-enterprise collaborative quality witness system in the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme of the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0050] The collaborative quality control method for preventing subjective human factor failure in nuclear power construction in the present specific embodiment includes the following steps:

[0051] (1) Based on the STAMP method, a subjective human error process model and a safety control structure are constructed for the collaborative quality witnessing process. The subjective human error process model is shown in FIG. 1, the formation path of subjective human error and safety control behavior of each participant in the collaborative quality witnessing process is analyzed, and multiple risk factors of subjective human error are identified according to the formation path of subjective human error and safety control behavior. Figure 1

[0052] (2) The DEMATEL-ISM method is used to calculate the importance ranking of each risk factor in preventing subjective human error, the correlation logic between risk factors is analyzed by constructing a multi-level hierarchical model of risk factors, and the risk factors are divided into organizational factors, team factors, individual factors and technical facility factors.

[0053] (3) The technical facility factor is selected for optimization and improvement, and a collaborative quality witnessing process and system architecture for preventing subjective human error are designed.

[0054] In the specific embodiment, in step (1), the subjective human error process model includes a controller, an actuator, a controlled object and a sensor. The actuator receives instructions and feedback information to make the controlled object controllable. The controller includes a general contractor, an owner and a supervision company. The actuator includes a witnessing supervised party. The controlled object includes a component or a process. The sensor includes a quality inspection instrument. According to the interaction between each control level, there are usually three basic control defects: the witnessing supervised party issues insufficient or improper control instructions; the witnessing supervised party does not perform the control action sufficiently; and the feedback information of the controlled process is lost or insufficient.

[0055] The safety control structure includes a management layer, an execution layer and a physical layer in order of importance. Each layer controls by imposing safety constraints on the lower layer subsystem, and the lower layer subsystem performs related operations under the safety constraints of the upper layer subsystem and feeds back the implementation information to the upper layer subsystem. Subjective human error events only occur when the safety constraints are violated or not successfully implemented. According to the specific safety constraints and information feedback activities of each layer in the collaborative quality witnessing, a hierarchical control structure for preventing subjective human error is constructed, as shown in FIG. 2. Figure 2

[0056] ​​The risk factors classification in the subjective human error prevention hierarchical control structure is improved from the behavioral science perspective, which is divided into organizational factors, team factors, personal factors and technical facility factors of subjective human error causes. The correlation and safety constraints of each level are analyzed to provide the basis for identifying potential causal factors of subjective human error. The unsafe control behaviors of the management layer in the first level of the control structure correspond to the organizational factors of subjective human error causes, while the three-level quality supervision of the execution layer and the operating personnel correspond to the team factors and the personal factors, and the physical layer of the instrument equipment provides information feedback corresponding to the technical facility factors.

[0057] According to the subjective human error process model and the failure prevention hierarchical control structure of collaborative quality witness, the non-safe control behaviors caused by basic control defects in each system level are analyzed, and the risk factors leading to subjective human error events are identified from the perspective of behavioral science. A preliminary analysis obtains 20 risk factors of subjective human error in the nuclear power equipment construction process. After classification and summary combined with investigation and expert opinions, 12 risk factors are retained. The system level of the risk factor category and the specific risk factors are shown in Table 1:

[0058]

[0059]

[0060] Table 1

[0061] In the specific embodiment, in step (2), when calculating and analyzing the importance and correlation logic of the risk factors, the following steps are adopted:

[0062] (a1) Determine the influence degree value between each factor. Use the Delphi method to score the risk factors. When i=j, α ij =0, the direct influence matrix A' is obtained; use the five-granularity fuzzy evaluation language set to replace the scores in the direct influence A' with triangular fuzzy numbers. Table 2 is the triangular fuzzy number conversion table,

[0063]

[0064] Table 2

[0065] Use the barycenter method to de-fuzzify the triangular fuzzy numbers, such as the formula:

[0066]

[0067] The converted direct influence matrix B' is obtained. Repeat the above method to score the risk factors for 10 times to obtain 10 converted direct influence matrices B'. Take the average of the 10 converted direct influence matrices B' to obtain the average matrix B.

[0068] (a2) Normalize the average matrix B to get the standard influence matrix C, as shown in Equation 1:

[0069]

[0070] Calculate the comprehensive influence matrix T, T = t ij , as shown in Equation 2:

[0071] T = C(I - C) -1 (2)

[0072] Determine the indirect influence relationship between each risk factor, and then calculate the influence degree r of the risk factor i , as shown in Equation 3:

[0073]

[0074] Affected degree s i , as shown in Equation 4:

[0075]

[0076] Centrality m i , as shown in Equation 5:

[0077] m i = r i + s i (5)

[0078] Reason degree n i , as shown in Equation 6:

[0079] n i = r i - s i (6)

[0080] To determine the importance and properties of each risk factor, as shown in Table 3:

[0081]

[0082] Table 3

[0083] (a3) Calculate the overall influence relationship matrix H, where I is the unit matrix, H = I + T, establish the reachable matrix K, and give the threshold value λ, as shown in Equation 7:

[0084]

[0085] k ij = 1 indicates that factor a i can influence factor a j , k ij = 0 indicates that factor a i cannot influence factor a j, the threshold value λ is added according to the mean value α and the standard deviation β of all elements in the comprehensive influence matrix T;

[0086] According to the hierarchical division of the reachable matrix K, the reachable set P i and the predecessor set Q i are first determined, the intersection of the two is U i , and the hierarchical division is performed by using the result priority extraction method. If P i = U i , the element is taken as the first level, and the corresponding row and column in the reachable matrix K are deleted, and the above process is repeated until all risk factors are allocated. The risk factor multi-level hierarchical model is drawn as shown in Figure 3 , and the correlation logic between the risk factors is analyzed from the risk factor multi-level hierarchical model.

[0087] In the specific embodiment, in step (3), a collaborative quality witnessing process for preventing subjective human factor failure is established based on Internet of Things technology and big data technology, as shown in Figure 4 , the collaborative quality witnessing process includes the following steps:

[0088] (b1) The operation personnel complete the collection of equipment, personnel and environment data through the handheld terminal device, and realize real-time intelligent perception and interconnection of witnessing process data information through Internet of Things technology access;

[0089] (b2) Based on the processing, storage, analysis and mining capacity of big data, the data of user qualification audit, personnel identity authentication, quality photos, video records and test report of process are cleaned, dimensionally reduced and mined, and the equipment quality state, witnessing execution progress and personnel operation behavior of the nuclear power construction process are identified;

[0090] (b3) Establish a quality intelligent witnessing platform, complete the quality witnessing management, quality test management and system basic management functions of the quality intelligent witnessing platform, integrate the witnessing process quality data, and display and interact in the quality intelligent witnessing platform in a visual way, and identify, supervise and control the subjective human factor failure risk in the nuclear power equipment quality witnessing process.

[0091] In the specific embodiment, in step (3), the core values of supporting the value chain collaboration of nuclear power equipment, nuclear safety culture, nuclear power laws and regulations, and quality assurance system are taken as the operation guarantee of collaborative quality witnessing, process control is taken as the goal, and the system architecture of the cross-enterprise collaborative quality witnessing system is designed. The system architecture includes an edge perception layer, a center processing layer and an application service layer, as shown in Figure 5 ;

[0092] The edge perception layer includes physical resource elements and data collection, the physical resource elements include nuclear power items, quality test procedures, and quality witness personnel; the data collection collects objective data and information of physical resource element situations in real time through sensing devices, quality test instruments, and video monitoring, the objective data and information of physical resource element situations include basic attributes and location information of items, process requirements of procedures, detection results of quality, identity information and human-computer interaction behaviors of quality state personnel;

[0093] The center processing layer converts information data used and interacted by the witness business in nuclear power construction into standardized data through cleaning, integration, and dimensionality reduction processing by big data technology, completes data specification packaging through natural semantic processing and language pre-training model analysis and mining, completes real-time transmission based on data standard protocol, forms end-to-end cross-enterprise information docking and is finally integrated in a unified data space;

[0094] The application service layer provides intelligent services supporting collaborative quality witness management business with process control as the goal, the application service layer includes business process improvement, knowledge push optimization, execution personnel supervision, and file security service;

[0095] Business process improvement is carried out around the electronic business process of collaborative quality witness, and promotes the standardization, systematization, and intelligentization of collaborative quality witness;

[0096] Knowledge push optimization provides procedures and experience knowledge of collaborative quality witness in line with legal regulations and technical route requirements;

[0097] Execution personnel supervision is to confirm and supervise quality witness personnel;

[0098] File security service provides quality process files, archival file records, and effectiveness verification of collaborative quality witness, and provides basis for quality problem traceability.

[0099] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting, although the present application has been described by referring to the preferred embodiments of the present application, those skilled in the art should understand that various changes can be made in form and details without departing from the spirit and scope of the present application defined by the appended claims.

Claims

1. A collaborative quality control method for preventing human subjective failure in nuclear power plant construction, characterized by: Comprising the following steps: (1) Constructing a subjective human error process model and a safety control structure based on the STAMP method, analyzing the formation path of subjective human error and safety control behavior of each participant in the collaborative quality witnessing process, and identifying multiple risk factors of subjective human error according to the formation path of subjective human error and safety control behavior; (2) Calculating the importance ranking of each risk factor in preventing subjective human error by using the DEMATEL-ISM method, analyzing the correlation logic between risk factors by constructing a multi-level hierarchical model of risk factors, and dividing the risk factors into organizational factors, team factors, individual factors, and technical facility factors; In the calculation and analysis of the importance and correlation logic of risk factors, the following steps are adopted: (a1) determine the degree of influence between each factor, score n risk factors using the Delphi method, when i = j, there is a ij = 0, get the direct influence matrix A'; use the five granularity fuzzy evaluation language set, replace the score in the direct influence A' with a triangular fuzzy number, use the barycenter method to de-fuzzify the triangular fuzzy number, get the transformed direct influence matrix B', repeat the above method to score the risk factors multiple times to get multiple transformed direct influence matrices B', and take the average of the multiple transformed direct influence matrices B' to get the average matrix B; (a2) Standardizing the average matrix B to obtain the standard influence matrix C, as shown in formula 1: Compute the comprehensive influence matrix T, T = t ij As equation 2: T = C(I - C) -1 (2) Determine the indirect influence relationship between each risk factor, and then calculate the influence degree r of the risk factor i As formula 3: Degree of influence s i As Formula 4: Centrality m i As equation 5: m i = r i + s i (5) Reasoning degree n i As Equation 6: n i = r i - s i (6) To determine the importance and attributes of each risk factor; (a3) Calculating the overall influence relationship matrix H, where I is the unit matrix, H = I + T, establishing the reachable matrix K, and giving a threshold value λ, as shown in formula 7: k ij = 1 indicates that factor a i can be influenced by factor a j , k ij = 0 indicates that factor a i cannot be influenced by factor a j , the threshold value λ is added according to the mean α and the standard deviation β of all elements in the comprehensive influence matrix T According to the reachable matrix K, the hierarchical division is performed by determining a reachable set P i and an antecedent set Q i , and the intersection of the two sets is U i . The hierarchical division is performed by using a result-priority extraction method. If P i = U i , the factor is taken as the first level, and the corresponding row and column are deleted from the reachable matrix K. The above process is repeated until all risk factors are assigned. A multi-level hierarchical model of risk factors is drawn, and the correlation logic between the risk factors is analyzed from the multi-level hierarchical model of risk factors. (3) Selecting technical facility factors for optimization and improvement, designing a collaborative quality witnessing process and system architecture to prevent subjective human error.

2. The collaborative quality control method for preventing human factor failures in nuclear power plant construction of claim 1, wherein: In step (1), the subjective human error process model includes a controller, an actuator, a controlled object, and a sensor. The actuator receives instructions and feedback information to make the controlled object controllable. The controller includes a general contractor, an owner, and a supervision company. The actuator includes a witnessing supervised party. The controlled object includes parts or processes. The sensor includes a quality inspection instrument. The safety control structure includes management layer, execution layer and physical layer in order of primary and secondary, each layer controls by imposing safety constraints on the lower subsystem, while the lower subsystem operates under the safety constraints of the upper subsystem and feeds back the implementation information to the upper subsystem.

3. The collaborative quality control method for preventing human factor failures in nuclear power plant construction of claim 1, wherein: In step (3), the collaborative quality witnessing process includes the following steps: (b1) The workers complete the collection of equipment, personnel and environment data through handheld terminal devices, and realize real-time intelligent sensing and interconnection of witnessing process data information through Internet of Things technology; (b2) Based on the processing, storage, analysis and mining capabilities of big data, the data of user qualification review, personnel identity authentication, quality photos, video records and test reports generated in the collaborative quality witnessing process are cleaned, dimensionally reduced and mined to identify the equipment quality status, witnessing execution progress and personnel operation behavior in the nuclear power construction process; (b3) Establishing a quality intelligent witnessing platform, completing the quality witnessing management, quality test management and system basic management functions of the quality intelligent witnessing platform, integrating the quality data of the witnessing process, and displaying and interacting in a visual manner in the quality intelligent witnessing platform to identify, supervise and control the subjective human error risk in the nuclear power equipment quality witnessing process.

4. The collaborative quality control method for preventing human factor failures in nuclear power plant construction of claim 1, wherein: In step (3), the system architecture includes an edge perception layer, a central processing layer and an application service layer. The edge perception layer includes physical resource elements and data collection, the physical resource elements including nuclear power items, quality test procedures, and quality witness personnel; the data collection collecting objective data and information of physical resource element situations in real time through sensing devices, quality test instruments, and video monitoring, the objective data and information of physical resource element situations including basic attributes and location information of items, process requirements of procedures, detection results of quality, identity information and human-computer interaction behaviors of quality status personnel; The center processing layer converts information data used and interacted in the witnessing business in nuclear power construction into standardized data through cleaning, integration, and dimension reduction processing by big data technology, completes data specification packaging through natural semantic processing and language pre-training model analysis and mining, completes real-time transmission based on data standard protocols, forms end-to-end cross-enterprise information docking, and is finally integrated in a unified data space; The application service layer provides intelligent services supporting collaborative quality witnessing management business with process control as the goal, the application service layer including business process improvement, knowledge push optimization, execution personnel supervision, and file security service; The business process improvement is developed around the electronic business process of collaborative quality witnessing, and promotes the standardization, systematization, and intelligentization of collaborative quality witnessing; The knowledge push optimization provides procedures and experience knowledge of collaborative quality witnessing in line with legal regulations and technical route requirements; The execution personnel supervision is the confirmation and supervision of quality witnessing personnel; The file security service provides quality process files, archival file records, and effectiveness verification of collaborative quality witnessing, and provides a basis for the traceability of quality problems.

Images