Europe User Conference: BP - implementing and supporting a multi-model, multi-site software roll-out
1 like197 views
This document summarizes BP's program to implement a standardized process modeling software, Petro-SIM, across its global refineries. The program aims to (1) select a single simulation platform, (2) develop standard process models, and (3) establish support roles to improve efficiency. Key aspects include evaluating Petro-SIM and models, phased rollout, and providing ongoing training and technical support. The program has rolled out several validated models successfully but also faces challenges regarding user workflows and software limitations that require customization or future releases.
Europe User Conference: BP - implementing and supporting a multi-model, multi-site software roll-out
- 1. Implementing and supporting a multi-model, multi-site software roll-out Refining Technology & Engineering
- 2. Introduction • Provide a high level summary of the program for the deployment of Petro-SIM and the process models within the BP refining business. • A program born from the need and desire to standardise the process modelling tools in use • The detailed program covers – Model Evaluations – Roll-out to refineries – Models support ©BP 2018 2 Standard, high quality UI KBC reactor models BP IP KBC Petrosim
- 3. Context • Process Models are widely used within BP refining business – Process Monitoring – Unit troubleshooting – Project development – What-if analysis – Commercial tools updating and assurance • Diverse portfolio of refineries – Different history – Different levels of users’ proficiency in process modelling – Geographic and cultural differences – Combination of in-house and third-party applications ©BP 2018 • Objective – Single simulation platform for all refineries – Standard refining simulation process models – Establish a common operating model – Improve efficiency – Utilise resources more effectively to support and sustain the use of the process modelling tools 3
- 4. Expected Value • Preserve current level of benefits brought by process modelling • Improve efficiency and capabilities within the process modelling team and across the refineries • Shift focus of the modelling resources team to the creation of additional value • Enable modelling and optimization of larger refinery systems and entire refineries ©BP 2018 4 SustainabilityValueEfficiency
- 5. Program Outline • Vendor Selection – Evaluation of several vendor platforms – Selected Petro-SIM for a detailed evaluation and Pilot study • Phased Process Models Roll-out to all BP refineries – Detailed Evaluation of each process model against documented requirements gathered from current users and process experts. This included comparison against pilot plant data, well accepted internal models and commercial units – Roll-out to refineries • Ongoing Support ©BP 2018 5 201920152014 2016 20182017 Platforms Evaluation Models Roll-out
- 6. Operating Model ©BP 2018 6 IT Central Process Modelling Team Site Model Owner Site Model User KBC Support HardwareandLicensingsupport ProcessModellingcompetence andcapabilitydevelopment CENTRALBUSINESS KBC HW and licensing support Central Engineering Team Software Hardware Build and Maintain model How we use the model Consumers of tools and information
- 7. Roles and responsibilities • Central Teams – Process Modelling Configures and deploy the process models Drive the use of the process models Provide process modelling expertise throughout BP refining – Engineering Establish best practices for unit performance test Support sites to close data quality gaps Use the models as part of their role – IT SW distribution Computer support (installation and licenses issues, HW and SW issues) Support model database • Refineries – Site leadership Drives model adoption across refinery – Site model owners Ensures accuracy of mass balance and other process data Keep model up to date – Site model users ©BP 2018 7
- 8. Process Model Evaluation Before being rolled-out each process model is subject to a detailed evaluation to ensure it is fit for purpose and that its roll-out does not result in performances “back-sliding” compared to existing tools. • Detailed evaluation of each process model against: – Functional Design Specification – Pilot Plant Data – Commercial Unit data – Existing modelling tools in use • Gaps identification – Gaps are identified, documented and prioritised – Regular meetings are held with KBC during the evaluation to discuss the identified gaps, to work towards a resolution or to identify temporary workarounds – Critical gaps prevent the model from being rolled-out • Gaps resolution – Work closely with KBC to close gaps and roll out updated models via most appropriate channels (e.g. hot fixes, service Packs, inclusion in software development path) ©BP 2018 8
- 9. Roll-out to refineries Standard work process for model roll- out to refineries • Data Collection • Model configuration and calibration • Users training • Model hand-over to users • User acceptance process • Model support ©BP 2018 9
- 10. Process Model Structure ©BP 2018 10 Excel Model Interface • Designed and supported by PM Central Team • Used For • Unit monitoring • Activity Tracking • LP Generation • Troubleshooting • Used by • Majority of model Users Flowsheet Base Case • Configured and tuned by PM central team and Model owner • Used For • Detailed Tuning • H&M balance • Equipment and What- if studies • Used by • Expert users • Project Users Model Inputs Model Inputs Model Outputs
- 11. Main Deliverables • Petro-SIM simulation platform for general simulation processes • Configured and calibrated models • Excel model interface and integration with data work-up tool for data transfer between data historian and KBC workbook • Capability for using models • Integrated support from RTE & IT&S • Database for calibrations sharing – Long term deliverable not included in the initial roll-out ©BP 2018 11
- 12. User Acceptance Criteria Each delivered process model goes through an acceptance process which covers several criteria: • Data quality and data handling functionalities • Model configuration – Does the model represent all the required equipment, streams, properties, operating conditions? – Are operating modes represented? • Model performance – Products yields and products qualities vs. plant data – Reactor performance vs. plant data – Catalyst/unit performance tracking – Step-out case against key process variables • Model robustness • Once accepted the model is owned by the refinery and is the single tool for all use cases ©BP 2018 12
- 13. Model Life ©BP 2018 13 ModelOwner CentralPM teamModelUser Model Update Needed Configuration update (if required) Data Selection for calibration Model Calibration Model Published Download New model Initiate regular model assuranc e process Review model performance Model Fit for purpose? KBC Document gaps due to bugs or model development Review bugs/model improvements Agree timeline for fixes and improvements Agree timeline for fixes and improvements No
- 14. Program Management • BP-KBC monthly and weekly meeting – Review status of open items for reactor models and Petro-SIM – Review status of development activities – Short-term resolution (Service Pack) – Future releases • Regular BP-KBC meeting during evaluation of each process model • Regular BP internal meetings – Track and report program progress – Coordinate model rollout plans against (ever-changing) refinery time & resource constraints – Stakeholder and management communication • Weekly meetings between Process Modelling Engineers and Site Process Engineers throughout model roll-out ©BP 2018 14
- 15. Successes to date • Several models rolled-out to the refineries after successful evaluation • Models rolled out so far well received by users • Excel interface appreciated for its ease of configuration and flexibility • Very positive relationship established with KBC to address gaps resolutions, bug fixes, and general support during model implementation ©BP 2018 15
- 16. Challenges • Work process often perceived by new users as complex • Some non-standard unit configurations required customization of the Base case and workbook • For complex units speed of execution can be an issue • Tuning of ULSD mode for DHTR-Sim complex • Some Process model in their current state haven’t passed the evaluation phase and will be re-evaluated in new releases ©BP 2018 16