A partial observable petri net based fault diagnosis and optimal sensor placement method for high-speed railway signal system

Abstract

The application discloses a partial observable Petri net fault diagnosis and optimal sensor configuration method of a high-speed rail signal system, and comprises the following steps: step 1, establishing a label Petri net model of the high-speed rail communication signal system; step S2, constructing a corresponding proof net model of the label Petri net system; step S3, constructing basic bad paths of the label Petri net system; the partial observable Petri net fault diagnosis and optimal sensor configuration method of the high-speed rail signal system; step S4, converting each basic bad path into corresponding inequality constraint conditions; step S5, applying a related program package to a corresponding integer linear programming problem to obtain an optimal solution of transition re-marking. The application fuses the proof net and system structure analysis technology, and the analysis process does not need to enumerate the entire state space, so that the calculation complexity of the algorithm is reduced.

Description

Technical Field

[0001] This invention relates to the field of information systems technology, and in particular to a method for diagnostic analysis of partially observable Petri nets and optimal sensor configuration in high-speed rail signaling systems. Background Technology

[0002] With the increasing speed and density of high-speed trains, efficient and safe signaling systems have become crucial for ensuring the safe operation of railways. Natural disasters and accidents pose a significant threat to the safe operation of high-speed trains. Therefore, ensuring the high precision, fast transmission, and strong anti-interference capabilities of the high-speed train communication and signaling system is key to improving the safety of high-speed train communication systems.

[0003] Current research and fault diagnosis of high-speed rail communication and signaling systems mainly focus on fault diagnosis of subsystems or signaling equipment. This paper combines the structural characteristics of the labeled Petri net model with key communication nodes in the high-speed rail signal transmission process, systematically studying the conditions for system diagnosability and the optimal sensor configuration problem for monitoring system signal transmission. A special network system calibration network capable of mining the diagnosability conditions of the Petri net is constructed. Based on the structural characteristics of the calibration network, a method for reconstructing the optimal label function is proposed, and conventionally identifiable transitions are relabeled and reassigned. The label function selection problem is transformed into the optimization problem of the additional sensor cost for monitoring high-speed rail signal transmission or processing, and is handled using an integer linear programming (ILPP) solution. Summary of the Invention

[0004] The communication and signaling system of high-speed railways is an automated system that completes train operation control and management. It mainly consists of three subsystems: the Train Control System (CTCS), the Interlocking System (CBI), and the Centralized Dispatch Center (CTC), along with a signaling auxiliary system, forming a complex distributed system. These subsystems are both independent and closely interconnected. Due to the aforementioned issues, the high-speed railway signaling system contains unobservable states and unidentified signal transmission nodes. To ensure the diagnosability of the Petri net model of the high-speed railway signaling system, this invention provides a method for partially observable Petri net diagnostic analysis and optimal sensor configuration for high-speed railway signaling systems. By constructing a partially tagged Petri net model of the high-speed railway communication and signaling system, the method proposed in this patent is used to verify the model's diagnosability, and the optimal sensor configuration scheme for model diagnosis is calculated, thereby improving the accuracy, transmission speed, and anti-interference capability of the high-speed railway communication and signaling system.

[0005] Technical solution: To achieve the above objectives, the technical solution adopted by this invention is as follows:

[0006] A method for partial observable Petri net fault diagnosis and optimal sensor configuration in high-speed rail signaling systems includes the following steps:

[0007] Step 1: Establish a Petri net model for the high-speed rail communication and signaling system.

[0008] Step 2: Construct the proofreading network model corresponding to the Petri net system;

[0009] Step 3: Construct the basic bad paths of the labeled Petri net system;

[0010] Step 4: Transform each basic bad path into a corresponding inequality constraint.

[0011] Step 5: Apply the relevant program packages to the corresponding integer linear programming problem to obtain the optimal solution for transition relabeling.

[0012] Beneficial effects: This paper proposes a label function reconstruction algorithm that links the structural characteristics of the system with its diagnosability. This method does not require iterative computation, avoids converting the system into a finite state automaton, and is applicable to most bounded and unbounded Petri net systems.

[0013] A tagged Petri net model is constructed based on the communication mechanism of the high-speed rail communication signaling system. The diagnosability of the system is analyzed, and the optimal configuration scheme of the sensors that the model can diagnose is calculated.

[0014] By integrating the calibration network with system structure analysis technology, the analysis process does not need to enumerate the entire state space, thus reducing the computational complexity of the algorithm. Attached image description:

[0015] Figure 1 This is a Petri net model diagram of the high-speed rail signaling system.

[0016] Figure 2 This is the proofreading network diagram corresponding to the labeled Petri net model.

[0017] Figure 3 This is the basic bad path graph of the labeled Petri net system. Specific implementation methods

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Two algorithms related to this invention are introduced. The first algorithm is:

[0020] 1. Proofreading network construction algorithm

[0021] Input for proofreading: Petri net

[0022] Output of the proofreading network: Label proofreading network system

[0023] Step 1: Establish a subset of induced transitions of N

[0024] Step 2: For all transitions t f ∈T f Add a transition. Represented as (λ,t) f For all p∈P′, make For all p∈P, make

[0025] Step 3: For all t reg ∈T reg Add a transition. Represented as (λ,t) reg For all p∈P′, we have For all p∈P, we have

[0026] Step 4: For all t′ reg ∈T′ reg Add a transition. Represented as (t) r ′ eg For all p∈P′, we have

[0027] Step 5: For all transitions t with labels l∈L, t′0 and t0 satisfy t′0∈T′0, t0∈T0.

[0028] Add a transition Represented as (t0′, t0); for all p∈P′, we have For all p∈P, we have

[0029] For the induced transition subset of N Let: T′=T / T f =T o ∪T reg ;

[0030] For the input: N = {P, T, Pre, Post}, T = T o ∪T u ∪T f , T→L∪{ε};

[0031] For the output:

[0032] For the induced transition subset of N Let: T′=T / T f =T o ∪T reg ;

[0033] 2. ILPP Method for Integer Linear Programming Problems with Sensor Configuration

[0034] Input: Petri net with given labels

[0035] Output: The proofreading network system corresponding to the labeled Petri net.

[0036] Step 1: Capture the VN system The bad paths are identified, and only transitions or transition pairs that affect the system's diagnosability are retained, resulting in all EBPs.

[0037] Step 2: Detect the transition pairs (γ′) in each EBP from the root node to the leaf node. i ,γ j Construct a set of linear algebraic constraints.

[0038] Step 3: Based on C1-C8, assign each transition pair (γ′) of EBP to... i ,γ j Transform into μ i inequalities, construct a set of linear algebraic constraints

[0039] C1, if γ′ i =t i ,γ j =t j If i ≠ j, then add +v. i +v j ;

[0040] C2, if γ′ i =λ,γ j ∈T r,o γ j =t i Then add +v i ;

[0041] C3, if γ′ i ∈T′ r,o γ j =λ,γ′i =t i Then add +v i ;

[0042] C4. If the transition (λ, γ) is continuous... i )(γ′ i ,λ) or (γ′ i ,λ)(λ,γ i ), and γ i =t i , γ′ i =t i Then add +v i -v i or -v i +v i ;

[0043] C5, if γ i =t i ,γ j =t j If i = j, then no addition is needed;

[0044] C6. If γ i =λ,γ j =t f If so, no need to add;

[0045] C7. If γ′ i =λ,γ′ j ∈T r,uo If so, no need to add;

[0046] C8, if γ j =λ,γ′ i ∈T′ r,uo If so, no need to add;

[0047] Step 4: Calculate the linear algebra constraint set If a solution is found, the corresponding sensor configuration result will be output; otherwise, an error message will be output.

[0048] Based on the proposed label function construction method and transition relabeling algorithm, the diagnosability of the labeled Petri net of the high-speed rail communication and signaling system is analyzed, and the optimal selection method for system sensors is studied. First, a labeled Petri net model of the high-speed rail communication and signaling system is established. like Figure 1 As shown, the meanings of "repository" and "change" are given in Tables 1 and 2.

[0049] Table 1. Figure 1 The physical meaning of each warehouse in China

[0050] warehouse Meaning of warehouse warehouse Meaning of warehouse <![CDATA[P1]]> transponder and track circuit <![CDATA[P5]]> CTC extension <![CDATA[P2]]> CBI Interlocking System <![CDATA[P6]]> Train Control Center (TCC) <![CDATA[P3]]> Onboard ATP device <![CDATA[P7]]> CTC switchboard <![CDATA[P4]]> RBC <![CDATA[P8]]> TSRS

[0051] Table 2 Figure 1 Physical meaning of each change

[0052]

[0053] Depend on Figure 1 We know that T0 = {t1, t2, t3, t4, t6, t8}, T f ={t f1}, T r,o ={t5,t7,t9,t 10},

[0054] T r,o The transitions labeled ε are associated and denoted as t5, t7, t9, and t 10 This indicates that it can be re-recorded during application. This is a non-diagnostic system for ab(c). k For k≥0, observations can be made without including fault T. f The sequence t2t3(t7t5t4) k Proof, or by including fault t f1 The sequence t2t3t7(t9t) 10 t f1 t4t5) k or t2t3t7t9t 10 t f1 (t7t4t5) k Proof. Assume the high-speed rail communication signaling system satisfies assumption A3. If transitions t5 and t9 can be relabeled, i.e. Then the system can become a diagnostic system.

[0055] This embodiment proposes a method for partial observable Petri net fault diagnosis and optimal sensor configuration in a high-speed rail signaling system, including the following steps:

[0056] Step S1: Establish a tag Petri net model for the high-speed rail communication and signaling system.

[0057] Step S2: Expand the proofreading net.

[0058] T r,o The transitions labeled ε are associated and denoted as t5, t7, t9, and t 10 This indicates that it can be re-recorded during application.

[0059] Among them, by Figure 1 We know that T0 = {t1, t2, t3, t4, t6, t8}, T f ={t f1}, Tr,o ={t5,t7,t9,t 10},

[0060] Figure 1 Petri net system The intersection of the central repository P and the T′-induced subset repository P′ is the repository set of the collation network. The T′-induced subset is obtained by... Fault transition t f1 Obtained. The unified cost of the Petri net system is compared with T. o ∪T r,o Each transition in the model is interconnected; when running ILPP, the number of sensors can be reduced, while the addition or replacement of sensors alters the system's diagnostics. Observable transitions are represented by (l,l) and (t′0,t0), respectively. Figure 2 White squares are used to represent unobservable transitions, which are represented in only one way, such as (λ, t8). Since label a is related to two transitions (t1, t2), the collation net contains four transition labels (a, a), and the labeling method for label d can be obtained similarly. If there are reversible transitions between places, double arrows are used to represent them.

[0061] Step S3, given as follows Figure 1 The Petri net system shown has a corresponding calibration net VN as follows: Figure 2 As shown.

[0062] The RG (Regulatory Node) for establishing the calibration network (VN) contains 16 nodes. If only the expanded graph of the basic bad paths is considered, it contains only 10 bad paths, such as... Figure 2 As shown.

[0063] Step S4: According to Algorithm 2, Figure 3 Each basic bad path in the equation is transformed into a corresponding inequality constraint.

[0064] In equation (3), (3.1)-(3.10) correspond to 10 basic bad paths in the system.

[0065]

[0066] EBP(3.1):

[0067] According to the C6 transition (λ,t) f No additional constraints are added.

[0068] According to C1, the transition (t1, t′2) corresponds to v1 + v2;

[0069] According to C3, (λ,t9) and (λ,t) 10 Add +v9 and +v to (λ,t7) respectively.10 +v7.

[0070] EBP(3.2):

[0071] According to C1, the transition (t1, t′2) corresponds to v1 + v2;

[0072] According to C3, transitions (λ,t5), (λ,t7), and (λ,t5) correspond to +v5, +v7, and +v5, respectively; E

[0073] BP(3.3):

[0074] According to C1, transition (t1, t′2) corresponds to v1 + v2; transition (t′5, λ) corresponds to v5.

[0075] According to C5, transitions (t6,t′6) and (t8,t′8) do not add constraint terms;

[0076] EBP(3.4):

[0077] The transition (t1, t′2) corresponds to v1 + v2;

[0078] The transition (t′5,λ) corresponds to v5; both (t′6,t8) and (t6,t′8) correspond to v6+v8;

[0079] (λ,t9), (λ,t) 10 Add v9 and v to (λ,t7) respectively. 10 v7; Transition (λ,t) f1 No additional constraints are added.

[0080] EBP(3.5):

[0081] According to C5, the transition (t1,t′1) is not subject to any constraints, and both (t′6,t8) and (t6,t′8) correspond to v6+v8;

[0082] The transitions (t′5,λ) and (λ,t5) correspond to +v5 and +v5, respectively.

[0083] EBP(3.6):

[0084] The transition (t′1,t1) does not add any constraints.

[0085] The transition (t6, t′8) corresponds to v6 + v8;

[0086] Add +v7 and +v5 to (λ,t7) and (λ,t5) respectively.

[0087] EBP(3.7):

[0088] The transition pair (t1, t′2) corresponds to v1 + v2;

[0089] No constraints can be added to the transition (t′3,t3);

[0090] The transitions to (λ,t7), (λ,t5), and (t′5,λ) are respectively modified by adding +v5, +v7, and +v5;

[0091] The transition pairs (t′6,t8) and (t6,t′8) both correspond to v6+v8.

[0092] EBP(3.8):

[0093] No additional constraints are added to the transitions (t′3,t3) and (t′4,t4);

[0094] The transition pair (t1, t′2) corresponds to v1 + v2;

[0095] The transitions (λ,t7) and (λ,t5) correspond to +v7 and +v5, respectively.

[0096] EBP(3.9):

[0097] The transition (t1, t′2) corresponds to v1 + v2;

[0098] No additional constraints are added to the transitions (t′3,t3), (t′6,t6), and (t′8,t8);

[0099] The transitions (t′7,λ) and (t′5,λ) correspond to +v7 and +v5, respectively.

[0100] EBP(3.10):

[0101] The transition (t1, t′2) corresponds to v1 + v2;

[0102] The transition (t′3,t3) does not add any constraints.

[0103] The transitions (t′7,λ) and (t′5,λ) correspond to +v7 and +v5, respectively.

[0104] Add v6+v8 to the transition pair (t′6,t8).

[0105] Use the relevant packages to solve problems related to T new The constraints for all transitions are: And as shown in equation (3) c t This problem involves linear programming with integers = 1. Find the optimal solution for relabeling transitions. If relabeling is used... This enables the system to become a diagnostic system.

[0106] It should be noted that the above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations and substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for partial observable Petri net fault diagnosis and optimal sensor configuration in high-speed rail signaling systems, characterized in that, Includes the following steps: Step S1: Establish a tag Petri net model for the high-speed rail communication and signaling system. ; Step S2: Construct the proofreading network model corresponding to the labeled Petri net system; Step S3: Construct the basic bad paths of the labeled Petri net system; Step S4: Transform each basic bad path into a corresponding inequality constraint; Step S5: Apply the relevant program packages to the corresponding integer linear programming problem to obtain the optimal solution for transition relabeling; Step 2 involves constructing the proofreading network model corresponding to the labeled Petri net system, which is called unfolding the proofreading network; specifically, it includes the following steps: Input for proofreading: Petri net ;in, , , : ; Output of the proofreading network: Label proofreading network system :in, , : ; Step 2.1: Establish a subset of induced transitions of N. ,set up: ; ; ; Step 2.2: For all changes Add a transition , represented as For all ,make For all ,make , ; Step 2.3, for all Add a transition , represented as For all ,have For all ,have , ; Step 2.4, for all Add a transition , represented as For all ,have , ; Step 2.5: For all transition labels... The changes t, and satisfy ; In step 5, the relevant program packages are applied to the corresponding integer linear programming problem to obtain the optimal solution for transition relabeling, which specifically includes the following steps: Input: Petri net with given labels ; Output: The proofreading network system corresponding to the labeled Petri net. , ; Step 5.1: Capture the VN system The bad paths are identified, and only transitions or transition pairs that affect the system's diagnosability are retained to obtain all EBPs. Step 5.2: Detect transition pairs in each EBP from the root node to the leaf node. Construct a set of linear algebraic constraints ; Step 5.3: Based on C1-C8, convert each transition pair of EBP... Transform into about inequalities, construct a set of linear algebraic constraints : C1, if Then add ; C2, if , Then add ; C3, if , Then add ; C4. If there is continuous change or ,and , Then add or ; C5. If If so, no need to add; C6. If If so, no need to add; C7. If If so, no need to add; C8. If If so, no need to add; Step 5.4: Calculate the linear algebra constraint set. If a solution is found, the corresponding sensor configuration result will be output; otherwise, an error message will be output.

2. The method for partial observable Petri net fault diagnosis and optimal sensor configuration of a high-speed rail signaling system according to claim 1, characterized in that, In step 4, each basic bad path is transformed into a corresponding inequality constraint condition; as follows: 。

Images