Method for the generation of a set of conflicts for model-based system diagnostics, and corresponding diagnostic method

1. A computer-implemented method for generating a set of conflicts to diagnose a system describable by a model based on the arrangement of system components, said model including a set of primary relations (PRs) between inputs and outputs of each component, representing its operating function, and with which system a plurality of sensors is associated for the observation of variables indicative of operation conditions,

where said conflicts set includes a subset of system components for which the assumption of correct operation is not consistent with the observed system behaviour,

wherein the method includes the steps of:

generating, as a preliminary step, a complete set of Analytical Redundancy Relations (ARRs) for the system in implicit form, in which each relation represents the logical relations between a subset of observable variables of the system and a corresponding subset of support components of the system, and

for each diagnosis instance

performing a system simulation based on the model, given a predetermined set of values as system inputs, computing the expected values for a first subset of Analytical Redundancy Relations (D-ARRs) including the relations involving only one system observation variable,

comparing the expected and observed values of the system observation variables and identify the inconsistent observation variables, i.e. the unsatisfied redundancy relations of said first subset, indicating a fault in at least one system component; and

when discrepancies are detected,

checking the consistency of each relation belonging to a second subset of Analytical Redundancy Relations (I-ARR), comprising relations involving a plurality of system observation variables, at least one of which exhibits discrepancies between expected and observed values; and

identifying the support set of the unsatisfied Redundancy Relations of said second subset as set of system conflicts, adapted to detect and isolate the system faults.

2. Method according to claim 1 , characterized in that it includes the selection of a complete set of Analytical Redundancy Relations (ARR s ) for the system among a set of intermediate relations existing between observable and non-observable system variables,

wherein each intermediate relation (R i ) is generated by combining two predetermined relations

R j =( N j ,C j ,S j ,T j ) and R k =( N k ,C k ,S k ,T k )

N j , N k being the ordinal numbers assigned to the relations R j , R k , each being expressed in implicit form as tuple (I) of a subset S j of system variables, (II) of the set C j of the support components of said relation, and (III) of the set T j of the primary relations used to derive said intermediate relation,

according to the rule:

R =( N,C j ∪C k ,( S j ∪S k )−{ x},T j ∪T k ),

where (S j ∪S k )−{x} indicates a symmetric difference operation between the subsets of variables S j , S k for each non-observable variable xεS j ∩S k ,

when both the conditions:

S j ∩S k ≠{ } and T j ∩T k ={ }

are satisfied.

3. Method according to claim 2 , characterized by the fact that said set of intermediate relations does not include identical intermediate relations, according to the identity relation:

( S j ,T j )=( S k ,T k ).

4. Method according to claim 1 , wherein the consistency check of each relation of said second subset of analytical redundancy relations (I-ARR) is performed by checking the consistency of the results of the primary relations involved in the relation, expressed in implicit form.

5. Method according to claim 1 , comprising the operation of deriving the fault signature matrix for a single-fault theoretical condition, starting from the complete set of analytical redundancy relations generated in implicit form, to analyze the fault detection and isolation properties of the system and determine its ambiguity sets.

6. Method for the diagnosis of a system including a plurality of sensors observation variables representing operation conditions, said diagnosis being based on a system model describing the configuration of the system components and including a set of primary relations between inputs and outputs of each component, representing the corresponding operational function, said method including:

the system simulation based on said model;

the detection of system faults by comparison between the values expected from the simulation and the observed values of the system observation variables;

the generation of at least one conflicts set including a subset of system components for which the assumption of correct operation is not consistent with the observed system behaviour; and

the determination of the minimum number of components involved in the generation of conflicts between the expected system behaviour and the observed one, i.e. the intersection of all the generated conflicts;

characterized in that the generation of the conflicts set is carried out by a method according to claim 1 .

7. Diagnostic system, including a processing assembly programmed to execute a method for generating a conflicts set and to diagnose a system describable by a model, according to claim 1 .

8. Computer program or set of programs, executable by a processor, including one or more code modules to implement a method for generating a conflicts set and the diagnosis of a system describable by a model, according to claim 1 .