Summary of the invention
The object of the invention is the supervision transfer method of reliability decision providing a kind of nuclear power plant digitizing master-control room operator, do not set up corresponding supervision transfer fail-safe analysis program and mathematical model to solve the research of existing supervision displacement behavior, do not possess the technical matters on the basis of engineer applied.
For achieving the above object, the invention provides the supervision transfer method of reliability decision of a kind of nuclear power plant digitizing master-control room operator, comprise the following steps:
Step S1: be multiple supervision action by the procedure decomposition of operator to the supervision action of multiple man-machine interfaces of nuclear power plant's digitizing master-control room;
Step S2: judge described multiple time window monitoring action respectively;
Step S3: according to described multiple time window monitoring action, divides path and the node of the supervision transfer of described operator;
Step S4: the path of shifting according to the supervision of described operator and node, judges the described state monitoring transfer;
Step S5: according to the described state monitoring transfer, set up the Quantitative Analysis Model monitoring displacement behavior reliability;
Step S6: calculated by the Quantitative Analysis Model of described supervision displacement behavior reliability and monitor transfer probability of failure or monitor the transfer probability of success;
Step S7: according to described supervision transfer probability of failure or the supervision transfer probability of success, judges whether the supervision transfer of described nuclear power plant digitizing master-control room operator is reliable.
As a further improvement on the present invention:
The state of described supervision transfer comprises isomorphism and isomery, and in described step S5, the isomorphism markov that the Quantitative Analysis Model of described supervision displacement behavior reliability is the probability of failure of the supervision displacement behavior for isomorphism described in quantitative test shifts failure model and the isomery markov transfer failure model for the probability of failure of the supervision displacement behavior of isomery described in quantitative test.
The computing formula of described isomorphism markov transfer failure model is:
Wherein, S
jfor system state, H
ibehave because of state, A
kfor alarm condition, Rp is working specification state and Mq is Equations of The Second Kind management role state;
S
jt () is a jth system state, and when moment t, j=(0,1);
The state of Hi (t) residing for i-th people, and when moment t, i=(0,1);
A
kt () is a kth state of reporting to the police, and when moment t, k=(0,1);
R
pt () is the state of p working specification, and when moment t, i=(0,1);
Mq (t) is the state of q two class shape management roles, and when moment t, k=(0,1);
T
k jrepresent the jth part in a kth functional block;
T
k irepresent i-th part in a kth functional block;
for people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, in K object element, operator monitor activity transfers to area information j from area information i;
for operator is at H
i, S
j, A
k, R
p, M
qunder state, monitor that K module occurs from i-th information transfer to the reliable probability of jth information;
for at H
i, S
j, A
k, R
p, M
qunder state, monitor a K module generation jth information transfer fault rate;
P{H
i(t) | H
i(t-1) for people is because of system state fault rate;
P{S
j(t) | S
j(t-1) } be power plant system state fault rate;
P{A
k(t) | A
k(t-1) } be alarm condition fault rate;
P{R
p(t) | R
p(t-1) } be working specification state fault rate;
P{M
q(t) | M
q(t-1) } be two class management role state fault rates.
Described system state j, alarm condition k, working specification state p and the fault rate of two class management role state q under the state status of correspondence are inquired about and are obtained from supervision Transfer Fault basic database.
The computing formula of described isomery markov transfer failure model is:
Wherein,
for at moment t, people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, transfer to the n-th component transfer process of jth block from m the component of i-th piece;
for people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, monitor a jth component of i-th piece;
for at moment t, people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, transfer to the failed probability of the n-th component transfer of jth block from m the component of i-th piece;
Q
j(t+ Δ t) is at moment (t+ Δ t), and people is because of state i, and system state j, alarm condition k, working specification state p, with two class management role state q, monitor the weight coefficient of i-th piece;
for people because of state i, system state j, alarm condition k, working specification state p, with two class management role state q, monitor a jth component probability of failure of i-th piece.
Described supervision transfer probability of failure comprises the probability of failure of the probability of failure of the supervision displacement behavior of isomorphism and the supervision displacement behavior of isomery;
The present invention has following beneficial effect:
The operator monitor transfer method of reliability decision of nuclear power plant of the present invention digitizing master-control room, first analyze the path monitoring transfer and node, Quantitative Analysis Model again by setting up reliability calculates and monitors transfer probability of failure, and then calculates the supervision transfer of described nuclear power plant digitizing master-control room operator is reliable.Monitor shift reliability interface data and computational tool for nuclear power plant digitizing master-control room operator human reliability analysis (HRA) and probabilistic safety assessment (PSA) provide; For digitizing master-control room operator monitor displacement behavior and fail-safe analysis thereof provide qualitative and quantitative method and instrument, monitor that Transfer Fault probability provides countermeasure for power plant reduces; And be the basis that raising nuclear power plant's digitizing master-control room operator monitor efficiency and performance establish engineer applied.
Except object described above, feature and advantage, the present invention also has other object, feature and advantage.Below with reference to figure, the present invention is further detailed explanation.
Embodiment
Below in conjunction with accompanying drawing, embodiments of the invention are described in detail, but the multitude of different ways that the present invention can be defined by the claims and cover is implemented.
Monitoring activity, as a part for operator's cognitive activities, is the information source of plant operators.Based on the Characteristic and law of operator's monitoring activity under DCS, surveillance operation is divided into two stages based on nuclear power plant's state from cognition, one is monitor target (information source to locking (the supervision identification object that current time is unique), i.e. DCS master-control room monitor unit or information area, object i) monitoring activity, this process is typical static cognitive activities, comprises monitor task confirmation, acquisition monitor message, monitor message identification, selects to monitor that strategy and supervision export; Subordinate phase monitors transfer, i.e. the supervision transfer of operator between target (information source), to complete information search or the dynamic process (as shown in Figure 1) monitoring path transfer.Based on power plant's current state and requirement, operator is more than iterative cycles two stage cognitive activities in monitoring process, to realize and dynamic supervisory and control continuous to power plant, for power plant's state estimation provides parameter and information.
Such as, see Fig. 1, the locking of operator to i-th information Objecti in the Object that nuclear power plant master-control room display device shows (information source) (Objectl, Object2...Objectn represent the 1st information source, the 2nd information source ... n-th information source) is discovered, and namely operator is to the acquisition cognitive process of locking information.The process monitoring activity of current information Objectl being transferred to the monitoring activity to next source of target information Object2 that in Fig. 1, operator shows from nuclear power plant master-control room display device, i.e. operator monitor transfer process.
Monitor the state of transfer:
There is same information module (display or screen in the supervision of operator to plant information, VDU) inner different aforementioned sources asks the transfer of (i.e. Ai, Bi, Ci and Di), be designated as isomorphism transfer, as in Fig. 2, the transfer activity other information source B1, C1 or D1 of same information module VDUl is transferred in monitoring point by operator from A1.
There is different information module (display or screen, VDU) and ask (i.e. VDU in the supervision of operator to plant information
1, VDU
iwith VDUn) transfer, be designated as isomery transfer, as in Fig. 2, operator by monitoring point from information module VDU
1transfer to other information modules VDU
ion transfer activity.
The monitoring activity of nuclear power plant operator is exactly the behavior of obtaining information from complexity dynamic working environment.Under normal or unusual service condition condition; although operator may by the impact of the importance of information on the supervision of information; but operator monitor activity launches under power plant's status information and operator's mental state model-driven, it is continuous a, dynamic process.From monitoring activity activity itself, in general, operator, to the transfer of the supervision of system state, normally according to the current state of system, and has nothing to do with the state before system.Although monitor that target has certain expection (particularly under accident/state-event), but monitor path and transfer process not expection property and obvious rule, have obvious randomness, therefore, monitoring process can be similar to and see stochastic process as.This class process without the version determined (Changing Pattern without inevitable), thus can not represent by accurate relationship, but can describe with random function.
Markov model is typically used for describing dynamic, continuous random process function, according to theoretical reasoning and experiment statistics analysis, can suppose to monitor that the whereabouts of transfer is only relevant with factor to the state in this monitoring point, therefore can suppose to monitor that transfer process is the continuous transfer process in time series of markov property, there is Markov, can simulate by Markov model.
Markov (Markov) model:
In a stochastic process, if at a time, relevant to the state of transition probability only with now residing of another kind of state by a kind of state development, and with this moment before residing state completely irrelevant, this process is called Markovian process.
Definition 1: establish { X
n, n ∈ N) and be an arbitrary sequence, N is time parameter, N={0,1,2,3......), S is state space, S={S
1, S
2... S
n, if n ∈ is N, so:
p(x
n=s
in|x
n1=s
il,x
n2=s
i2,....x
n-1=s
in-1}|=p(x
n=s
in|x
n-1=s
in-1} (1)
{ X
n, n ∈ N} is called Markov chain.Equation (1) means if be in S in moment n-m system state
in-1, so NextState n from n-2 to 1 with before time have nothing to do, be only S with current time n-1 state
in-1relevant.Briefly, if system state oneself know, system state in the future has nothing to do in the past with system, and this is just called markov feature.
Definition 2: establish { X
n, n ∈ N} is Markov chain, conditional probability: P{X
m+n=S
j| X
m=S
i,=P
ij(m, m+n)=P
ijn () is the n-th step transition probability of Markov chain.Especially, P
ijregard first step transition probability as.N walks transition probability, P
ijcan derive according to Chapm an-Kolmogorov equation, first step transition probability P
ijbe key point, transition probability is as follows:
This matrix has following two features:
1)0<=P
ij<=1 i,j∈{1,2,3......,N}
2)
i,j∈{1,2,3......,N)
See Fig. 6, the supervision transfer method of reliability decision of nuclear power plant digitizing master-control room operator of the present invention, comprises the following steps:
Step S1: be multiple supervision action by the procedure decomposition of operator to the supervision action of multiple man-machine interfaces of nuclear power plant's digitizing master-control room.
Step S2: judge multiple time window monitoring action respectively.
Step S3: the multiple time window monitoring action of root pick, the path of the supervision transfer of division operation person and node.
Step S4: the path of shifting according to the supervision of operator and node, judges the state monitoring transfer.
Step S5: according to the state monitoring transfer, sets up the Quantitative Analysis Model monitoring displacement behavior reliability.
Step S6: monitor transfer probability of failure by monitoring that the Quantitative Analysis Model of displacement behavior reliability calculates or monitor the transfer probability of success.
Step S7: according to supervision transfer probability of failure or the supervision transfer probability of success, judges whether the supervision transfer of described nuclear power plant digitizing master-control room operator is reliable.Can according to the numerical value of the prison transfer obtained depending on probability of failure (or by formula: monitor that the transfer probability of success=1-monitors that transfer probability of failure calculates prison transfer and looks the probability of success), inquire about various international and domestic industry standard (according to application scenario and the type of man-machine interface that relates to different, the standard judged is also different, the standard judged is determined according to practical situations, also can be that power plant is from the standard of establishing), can know and monitor transfer reliability value whether in allowed limits (the whether reliable project listed by standard of the supervision of operator transfer and criterion and determine) thereof, thus judge whether reliable.
Above-mentioned steps, by first analyzing the path and node that monitor transfer, then the Quantitative Analysis Model setting up reliability calculates supervision transfer probability of failure, and then the supervision transfer of Suan get nuclear power plant digitizing master-control room operator is reliable.Monitor shift reliability interface data and computational tool for nuclear power plant digitizing master-control room operator human reliability analysis (HRA) and probabilistic safety assessment (PSA) provide; For digitizing master-control room operator monitor displacement behavior and fail-safe analysis thereof provide qualitative and quantitative method and instrument, monitor that Transfer Fault probability provides countermeasure for power plant reduces; And be the basis that raising nuclear power plant's digitizing master-control room operator monitor efficiency and performance establish engineer applied.
Embodiment 1:
Below for the digitizing master-control room of the nuclear power plant shown in Fig. 3, specifically describe the present invention.
As shown in Figure 3, as DCS master-control room operator workstation has 5 computer monitors (VDU, the main monitor unit of operator) to be divided into corresponding monitoring module unit, VDU is designated as
1, VDU
2... VDU
5(monitoring target as 5).4 large LED screen be connected as a single entity (regard 4 LED as entirety, monitor target as 1) in operator workstation dead ahead are considered as 1 overall monitoring module, are designated as the 6th monitor unit LED.Communication (i.e. the world-of-mouth communication, communication system etc. of teams and groups) between operator teams and groups internal members is considered as 1 overall monitoring module, is designated as the 7th monitor unit OA.Operator obtains other means of plant information and mode (supervision etc. as the activity of local operation person strange land) divides as a whole monitoring module, is designated as the 8th monitor unit OtherParts (auxiliary unit of monitor task).Monitor that transfer refers generally to shift according to monitored information, the module transferred to is that the monitor message obtained according to the last time judges, thus determines where transfer to.
Based on operator monitor behavioral trait, object element module (i.e. 5 VDU of DCS master-control room operator workstation are monitored greatly with master-control room operator 8, big screen LED, crewmember's communication way and other parts be totally 8 unit) for monitor goal displacement region, consider in supervision transfer process and be mainly subject to system state, people is because of factor, warning system, 5 major influence factors such as two class management roles and station guide malfunction protocol (SOP), describe and the Markov model (see Fig. 4) building DCS nuclear power plant and monitor transfer process by Markov function.In Fig. 4, (appropriate section of Fig. 3 is asked for an interview in the description of intermediate module) each supervision object element has n partial information region, then following two kinds of modes (namely monitoring the state of transfer) realize by operator monitor transfer activity:
(1) operator monitor isomery transfer: operator monitor activity in the diagram 8 unit objects (No.1 ..., No.8) between the process of transfer.
(2) operator monitor isomorphism shifts: the different information area of operator monitor activity in each unit object of Fig. 4 or information source are (as INF
1, INF
2..., INF
n, wherein INF
irepresent i-th information area in object element) between transfer process.
Obviously, in nuclear power plant's digitizing master-control room operator monitor transfer Markov model, there is Markov chain two type of isomorphism Markov chain and isomery.
The supervision transfer method of reliability decision of the nuclear power plant digitizing master-control room operator of the present embodiment, comprises the following steps:
Step S1: be multiple supervision action by the procedure decomposition of operator to the supervision action of multiple man-machine interfaces of nuclear power plant's digitizing master-control room.
Step S2: judge multiple time window monitoring action respectively.
Step S3: according to multiple time window monitoring action, division operation person faces the path and node of looking transfer.
In practical application, the required information content (information content has referred to the supplementary required for a decision process) monitored in operator monitor process, by surveillance operation Task-decomposing (can knowledge based characterization method realize) with monitor that transfering node divides and obtain.The transfer path of relevant information, also obtains by decomposing monitor task.And the time window of operator's surveillance operation under power plant's transient state, accident condition (i.e. Looking Out Time section), can be obtained with code by PSA report.Alarm on average triggers probability, obtains by analog machine experiment statistics.
Step S4: the path of shifting according to the supervision of operator and node, judges to monitor that transfer is isomorphism transfer or isomery transfer.
Step S5: for isomorphism transfer or isomery transfer, sets up the isomorphism markov transfer failure model of the probability of failure of the supervision displacement behavior being used for quantitative test isomorphism and the isomery markov transfer failure model for the probability of failure of the supervision displacement behavior of quantitative test isomery respectively.
(1) the isomorphism markov transfer failure model of DCS nuclear power plant SNNP monitoring process:
By on the influence factor affecting operator monitor behavior, as system state (S
i), people is because of state (H
i), alarm condition (A
k), working specification state (R
p) and Equations of The Second Kind management role (M
q) state etc., the hidden Markov model of SNNP can be built, see Fig. 5.In Fig. 5, H
i(i=1,2..., n) represents that people is because of state, S
j(j=1,2 ..., m) represent system state, A
k(k=1,2 ..., s) represent alarm state, R
p(P=1,2..., y) represents working specification state, M
q(q=1,2 ..., x) represent two class management role states.The state value of aforesaid five aspect influence factors, and basic fault rate, the weight of often kind of influence factor each main composition factor, by monitoring that eye movement test obtains.Then the computing formula of isomorphism markov transfer failure model is:
Wherein, S
jfor system state, H
ibehave such as, because of state (state of people may have several, and, light state, generally, tense situation, high-pressure state, i represents the wherein a kind of state of people in certain moment), A
kfor alarm condition, R
pfor working specification state and M
qfor Equations of The Second Kind management role state.
S
jt () is a jth system state, and when moment t, j=(0,1);
The state (people because of state) of Hi (t) residing for i-th people, and when moment t, i=(0,1);
A
kt () is a kth state of reporting to the police, and when moment t, k=(0,1);
R
pt () is the state of p state of operation code, and when moment t, i=(0,1);
Mq (t) is q two class shape management roles, and when moment t, k=(0,1);
T
k jrepresent the jth part in a kth functional block;
T
k irepresent i-th part in a kth functional block;
for people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, in K object element, operator monitor activity transfers to area information j from area information i:
for operator is at H
i, S
j, A
k, R
p, M
qunder state, monitor that the fault rate of white i-th information transfer to jth information occurs K module;
for at H
i, S
j, A
k, R
p, M
qunder state, monitor a K module generation jth information transfer fault rate;
P{H
i(t) | H
i(t-1) for people is because of system state fault rate; The state of people is divided into: very nervous, nervous, light three kinds of states; Use H respectively
i(t)=2, H
i(t)=1, H
it ()=0 represents.
P{S
j(t) | S
j(t-1) } be power plant system state fault rate; Power plant system state generally comprises normal and abnormality, uses S respectively
i(t)=1, S
it ()=0 represents.
P{A
k(t) | A
k(t-1) } be alarm condition fault rate; If warning is normal, use A
kif t ()=0 represents. warning itself is not in triggering instance, but warning system have issued warning, so uses A
kif t ()=1 represents. warning system is in false alarm signal and have issued warning, so uses A
kt ()=20 represent.
P{R
p(t) | R
p(t-1) } be working specification state fault rate; Working specification state is divided into normal and abnormal, uses R respectively
i(t)=1, R
it ()=0 is shown not.
P{M
q(t) | M
q(t-1) } be two class management role state fault rates.Two generic tasks refer to when finishing the work, and can not complete this task by the soft interface of correspondence, at this moment need a series of soft interfaces opening other correspondence, and this series of soft interface is with regard to two generic tasks; The two generic task states called generally comprise normal and abnormal, use M respectively
i(t)=1, M
it ()=0 represents.
The fault rate basic under the state status of correspondence of described system state j, alarm condition k, working specification state p and two class management role state q is inquired about and is obtained from supervision Transfer Fault basic database.
(2) the isomery markov transfer failure model of nuclear power plant's monitoring process
Refer to the markov transfer process of isomery and monitor that displacement behavior occurs in the supervision transfer between different unit (as between 5 VDU, between VDU and big screen LED etc.), then the computing formula of isomery markov transfer failure model is:
Wherein,
for at moment t, people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, transfer to the n-th component transfer process of jth block from m the component of i-th piece;
for people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, monitor a jth component of i-th piece;
for at moment t, people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, transfer to the failed probability of the n-th component transfer of jth block from m the component of i-th piece;
Q
j(t+ Δ t) is at moment (t+ Δ t), people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, monitor the weight coefficient of i-th piece;
for people because of state i, system state j, alarm condition k, working specification state p, two class management role state q, monitor a jth component probability of failure of i-th piece.
Step S6: monitor transfer probability of failure by monitoring that the Quantitative Analysis Model of displacement behavior reliability calculates.
Step S7: according to supervision transfer probability of failure, judges whether the supervision transfer of described nuclear power plant digitizing master-control room operator is reliable.
Monitor that transfer probability of failure comprises the probability of failure of the probability of failure of the supervision displacement behavior of isomorphism and the supervision displacement behavior of isomery.
In summary, the present invention has the following advantages:
(1) qualitative and quantitative method and instrument can be provided for digitizing master-control room operator monitor displacement behavior and fail-safe analysis thereof, monitor that Transfer Fault probability provides countermeasure for power plant reduces.
(2) reliability interface data and computational tool can be monitored for nuclear power plant digitizing master-control room operator human reliability analysis (HRA) and probabilistic safety assessment (PSA) provide.
(3) can be the basis that raising nuclear power plant's digitizing master-control room operator monitor efficiency and performance establish engineer applied.
(4) enlightenment can be provided for other digitizing industrial system master-control room operator monitors transfer fail-safe analysis field, can be reduced it and monitor failure probability, improve operator's reliability, ensure the safety of industrial system.
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.