Remote Process Analysis for Process Analysis and Optimization

Transcription

1 Remote Process Analysis for Process Analysis and Optimization Gregory Shahnovsky Amir Or

2 The Challenges Ahead The worldwide oil and gas industry has undergone a major transformation in the last decade due to changes in regulatory and market forces, such as fluctuating crude prices, tighter regulation on product quality and emissions, shifting crude quality and fundamental changes in fuel demands At present, refineries must be flexible enough to respond immediately to crude oil changes and deviations in product demands as a result of the changing global economy. The required flexibility in the management and the complexity of the different processes, crude oils and distillates can only be achieved by stringent quality monitoring and process optimization

3 Refinery Industry Trends The modern refinery is a high-volume and generally low-margin enterprise. While revenue may be large, feedstock costs around 90% of the cash flow. After accounting for fixed and variable operating costs, the remaining margin is typically below 5% of revenues To increase the margin, the refiner must (1) reduce feedstock costs, (2) decrease operating expenses and (3) increase product revenue

4 Remote Analysis Technologies LPG Analyzer Locations Gas Plant Olefins Poly. Plant Gasoline Blending Gas Alky. Plant ic 4 Alkylate CDU Light Naphtha Reformer Light Naphtha Reformate FCC Gasoline Crude Oil After Preheater Heavy Naphtha Aromatic Extraction Aromatics Kerosene Kerosene / Diesel AGO Fuel Oil Hydrocracker Vacuum A Tower B C FCCU Diesel Blending Crude Oil Blending Coker Lubricant Oils Bitumen

5 Remote Analysis Technologies DCS Control Room Modbus 485 and/or TCP/IP Up to 3KM Standard communication Fibers-Optic 8 Field Units Remote PC (ANACON Client) Ethernet Modem The Analyzer located in a safe area Field Units located up to 3KM from the analyzer Field unit does not contains any electricity or valves Use standard communication fibers-optic

6 Field Unit Operation Conditions Haze free Max. Pressure psi (40 bar) Flow rate - 1 l/min to 3 l/min Temperature- Above Cloud point to +160 C (300 F) Field Unit Sample Cell Probe

7 Dependence Signal of Fiber Length UP TO 3 km NOT INFLUENCED BY FIBER LENGTH

8 Principles of Chemometrics Calibration Phase Reference Spectra Reference Measurements Calibration Model mix# HCl HF Unknown Spectra Measurement Phase Calibration Model A1(t) A2(t) tote changeover new lot changeover Wed 07/7 Fri 09/7 Sun 11/7 Tue 13/7 Thu 15/7 Sat 17/7 Mon 19/7 Wed Fri 21/7 23/7 Sun 25/7 Tue 27/7 Thu 29/7 Sat 31/7 Mon Wed 02/8 04/ Fri 06/8

9 The Bottleneck in NIR/FTIR Analytics The main drawback not related to the hardware but to finding the resources (time and money) to build the models and maintaining the model robustness. For the most economical incentive applications, such as CDU and blending, a successful implementation of NIR/FTIR can last up to 12 months. The bottleneck is not the NIR/FTIR spectrometry accuracy or the limits on laboratory resources, it is the fact that the model has to be robust. For this to happen the model has to cover all possible variations of the product composition, which is constantly changing over time. Therefore, although very powerful, current NIR systems are not plug and play devices. If changing crude influences the analyzer's behavior, the refinery operators will lose their confidence in the analyzer.

10 Successful Model Building Good Lab data and on Time Good Model Good prediction Bad Lab data non conforming with sample time GIGO Bad Model Bad prediction Importance of Time Matching

11 Freetune Software Solution Freetune is a software program composed of two sequential blocks. First, PLS regression is used to build a localized sensitive model (LSM), based on just a small number of samples. One of the characteristics of this model is its sensitivity to short term (hours) process variations. Since it is localized, the model can be quickly built and validated. This procedure takes up to one week and can be performed prior to startup. Second, a proprietary software program processes the model, together with the specified plant data, to accurately quantify the properties. This second part deals with predicting the long-term product variability.

12 Freetune Software Solution NIR Result using LSM System Clock (Time of measurement) Correction to NIR Result System Results History Laboratory Results + Sampling time NIR Result retrieval and Integrity check Correction parameters estimation NIR Spectra Laboratory Data Pre- Processing LSM Advanced Processing Monitored Variable Validity Flag

13 Crude Switching Problem Solving 45 R 2 = 0.91 crude change La b Va lu es Fre etun e R MSE P = 1.86 F (1 o C) R MSEP = 1.93 F ( 1.07 o C) R MSE P = F (0. 99 o C) Feb 08-Fe b 12-Fe b 16-Feb 19-Fe b 21-Fe b 26-Fe b 01-M ar 04-M ar 07-M ar 10-M ar 13-M ar 16-M ar 25-M ar 28-M ar 31-M ar Cloud Pt. ( o F) 03-Apr 06-Ap r 09-Ap r 12-Apr Cloud Point of Diesel Fuel

14 The Fusion Solutions Stream 1 Distributed Control System Stream 2 Stream 3 Sample System Discrete Analyzer Modbus Fiber optic cable NIR Analyzer with Freetune* software Response time: 5-10 min Accuracy: according to ASTM Availability: above 99% (full-redundant system) * Patent pending

15 ANACON Big Data Functionality ANACON Client Ethernet ANACON Ethernet Server (Modbus через RS-485, ANACON Client Process Analyzer Modbus, 4-20 ma, итд TCP/IP, Profibus) DCS/PI LIMS Process Analyzer Process Analyzer Virtual Analyzer Validation HMI Trends Reports Technical Support and Maintenance

16 ANACON Big Data Functionality SAP/LIMS Process Analysis DB Temperature Pressure Power Process monitoring Anomalies detection Operators training Equipment diagnostics Preventive actions Quality control Optimization Level Flow Corrosion

17 ANACON Big Data Functionality Multidimensional fusion of incoming data Multidimensional distribution of proper data Abnormality of novel events detection Clustering, decision trees, linear, polynomial, logistic regression, SVM Escalation of novelty real-time analysis, etc.

18 Conclusions Modern remote analysis technologies offers effective and proven solution for process quality control and optimization Analyzer management software enables data collection, verification, cross-validation and multidimensional fusion of incoming data for more effective process analysis and control The latest system release includes predictive data analysis, safety considerations, emission sources localization and equipment maintenance improvement, using linear programming and multidimensional distribution techniques Combination of remote process analysis with big data IIoT technology in oil and gas industries improves process monitoring, quality control, optimization and profitability

19 Thank You for Attention