Pipelines are the veins of the oil and gas industry. They are the most economical way todelivercrudeandrefinedproductsalonglargedistanceswithhugequantities,despite implyinglargecapitalinvestments. Pipelinesaresubjectedtomanydamagesorfailures with tremendous consequences for the business, community and for the environment. In thecontext of Pipeline Integrity, PipelineInspection Gauges(PIGs), that internally inspect the line, plays a key role. However, their usage is characterized by many challenges: 1) huge operational cost, 2) difficulties of their practical applications and 3) large time and effort required to analyse the collected data. Regarding the first two challenges, Eni developed the e-splora technology, which consists in low-cost sensors installed on low-cost PIG structure and collect data about various physical quantities, as pressure, temperature. This thesis addresses the third challenge and resolves two different important problems by processing PIG data. Since the e-splora collect data in time without knowing where a certain value has been recorded, the first problem is identifying the e-splora PIG location inside the pipeline at any time. The proposed solution is based on modelling the PIG movement inside the pipeline as a piston-like and exploiting the fluid flow rate data of the pipeline to estimate the progressive distance of the PIG. The method performance is assessed by comparing the measured pressure by the PIG with the hydrostatic pressure (fixed reference) using a modified Dynamic Time Warping algorithm. The methodology is applied to a case study of a real pipeline. The second problem is the automatic recognition of pipeline different geometries (flange, valve, etc.) from the PIG data: identifying through which element of the pipe the PIG is passing. So, by enabling this possibility, the pipeline geometries and other damages/anomalies can be detected. In this case, the thesis presents a novel framework based on Artificial Neural Networks classifiers that recognizes the specific geometry of the pipeline as a function of data-pattern collected by the PIG. The method has been applied to a case study of a test pipeline.
Gli oleodotti sono le vene dell’industria oil and gas. Sono il modo più economico per trasportare prodotti grezzi e raffinati su grandi distanze con grandi quantità, nonostante implichino ingenti investimenti di capitale. Le condutture sono soggette a molti danni o guasti con conseguenze enormi per l’azienda, la comunità e l’ambiente. NelcontestodiPipelineIntegrity,iPipelineInspectionGauges(PIGs),cheispezionano internamentelalinea,giocanounruolochiave. Tuttavia,illoroutilizzoècaratterizzato da molte sfide: 1) enormi costi operativi, 2) difficoltà delle loro applicazioni pratiche e 3) grandi tempi e sforzi richiesti per analizzare i dati raccolti. Per quanto riguarda le prime due sfide, Eni ha sviluppato la tecnologia e-splora, che consiste in sensori a basso costo installati su un PIG più economico e raccoglie dati su varie grandezze fisiche, come pressione e temperatura. Questa tesi affronta la terza sfida e risolve due diversi problemi importanti elaborando i dati del PIG. Poiché e-splora raccoglie i dati nel tempo senza sapere dove un certo valore è stato registrato, il primo problema è quello di identificare in qualsiasi momento la posizione del PIG e-splora all’interno della condotta. La soluzione proposta si basa sulla modellazione del moto del PIG all’interno della condotta come di un pistone e sullo sfruttamento dei dati di portata del fluido della condotta per stimare la distanza progressiva del PIG. Le prestazioni del metodo sono valutate confrontando la pressione misurata dal PIG con la pressione idrostatica (riferimento fisso) utilizzando un algoritmo di Dynamic Time Warping modificato. La metodologia viene applicata ad un caso di studio di una condotta reale. Il secondo problema è il riconoscimento automatico delle diverse geometrie della condotta (flangia, valvola, ecc.) dai dati del PIG. In questo modo, è possibile rilevare le geometrie della tubazione e altri danni/anomalie. Per questo caso, la tesi presenta un nuovo framework basato su classificatori di reti neurali artificiali che riconosce la geometria specifica della condotta in funzione dei dati raccolti dal PIG. Il metodo è stato applicato ad un caso di studio di una pipeline di test.
PIG data processing for oil and gas pipeline integrity assessment
TEODORI, MARCO
2018/2019
Abstract
Pipelines are the veins of the oil and gas industry. They are the most economical way todelivercrudeandrefinedproductsalonglargedistanceswithhugequantities,despite implyinglargecapitalinvestments. Pipelinesaresubjectedtomanydamagesorfailures with tremendous consequences for the business, community and for the environment. In thecontext of Pipeline Integrity, PipelineInspection Gauges(PIGs), that internally inspect the line, plays a key role. However, their usage is characterized by many challenges: 1) huge operational cost, 2) difficulties of their practical applications and 3) large time and effort required to analyse the collected data. Regarding the first two challenges, Eni developed the e-splora technology, which consists in low-cost sensors installed on low-cost PIG structure and collect data about various physical quantities, as pressure, temperature. This thesis addresses the third challenge and resolves two different important problems by processing PIG data. Since the e-splora collect data in time without knowing where a certain value has been recorded, the first problem is identifying the e-splora PIG location inside the pipeline at any time. The proposed solution is based on modelling the PIG movement inside the pipeline as a piston-like and exploiting the fluid flow rate data of the pipeline to estimate the progressive distance of the PIG. The method performance is assessed by comparing the measured pressure by the PIG with the hydrostatic pressure (fixed reference) using a modified Dynamic Time Warping algorithm. The methodology is applied to a case study of a real pipeline. The second problem is the automatic recognition of pipeline different geometries (flange, valve, etc.) from the PIG data: identifying through which element of the pipe the PIG is passing. So, by enabling this possibility, the pipeline geometries and other damages/anomalies can be detected. In this case, the thesis presents a novel framework based on Artificial Neural Networks classifiers that recognizes the specific geometry of the pipeline as a function of data-pattern collected by the PIG. The method has been applied to a case study of a test pipeline.File | Dimensione | Formato | |
---|---|---|---|
TEODORI_THESIS.pdf
non accessibile
Descrizione: testo della tesi (Teodori Marco)
Dimensione
22.13 MB
Formato
Adobe PDF
|
22.13 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/151260