A significant development in production optimization is the increasing availability of digital information in association with advancements in 4-D seismic, intelligent completions, reservoir modeling, and simulation and production flow monitoring and control. Managing such an environment requires the accumulation of information and data about the changes taking place and using it to adjust operating parameters to optimize production.

Applications and technologies developed in recent years process this information and data
Figure 1. Basic PRODML optimization use case. (Image courtesy of Energistics)  
and through adjustments in models or set points or other parameters, helping operators to optimize production. Operaters often employ a number of these applications and technologies in concert, but the applications do not interact and exchange data efficiently or effectively. In 2005, several major industry companies decided to work together to develop a framework that would solve this problem.

The result was PRODML, a collaborative initiative within the Energistics community of the energy industry intended to facilitate production optimization work processes by defining a set of standards, with the goal being a common language for the exchange and use of production data. This language is based on XML and Web Services technologies.

The founders of the PRODML initiative were BP, Chevron, ExxonMobil, Shell and Statoil. They were joined by solution providers Halliburton, Invensys, Köngsberg Intellifield, OSISoft, Petroleum Experts, Schlumberger, TietoEnator and Weatherford. Additions in 2007 were Aspentech, ConocoPhillips, Euriware, InfoSys, Matrikon, P2ES, Pioneer Natural Resources and TIBCO as well as Honeywell, IBM, Intelligent Agent, ONGC, Petris, Roxar, Satyam and Sensortran.

PRODML is a companion set of standards to the widely accepted WITSML, used to transfer drilling information. Maintaining alignment among PRODML, WITSML and other Energistics standards will help ensure that they can be used in concert for integrated use cases.

PRODML is used to allow oil and gas production software applications to exchange meaningful information as they work together to achieve optimization objectives. In the PRODML world, one application acts as a “client” and requests something from a second application – the “server.” The key to this exchange is that the client must know how to ask for the information, and the server must understand the question and be able to create and return an answer to the client. Once the client receives the reply, it must be able to understand the answer.

Data object specifications

The data object specifications (i.e., XML Schemas) used in PRODML are derived from previously defined objects that use the architecture and style of the companion WITSML standards. While WITSML uses Web Services exclusively for adding and retrieving data from data transfer servers, PRODML uses Web Services interfaces to support a variety of application-to-application interactions.

PRODML (and WITSML) data specifications consist of a set of independent but related data object schemas. The data object schemas define sets of data that can be transmitted within single XML documents as cohesive subsets (e.g., production measurements, flow network, well test) of an overall logical information model related to a single domain. Data object schemas contain attributes, elements and included component sub-schemas.

In order to ensure that PRODML-based applications work together, PRODML defines the method for data exchange (that is, the Web Services interfaces and associated methods), the valid mechanisms for asking for information (with queries based on the schemas), and valid formats and context for the answers provided by a server (XML data schema made of up of predefined objects – the production data object schemas).

The data information model defined by PRODML allows information to be exchanged along with context information that describes the data. For example, when an application sends production volume values, it includes information that defines the well for which the data is provided and also date-time stamps for the associated values.

By defining an agreed-upon set of data object schemas, PRODML client applications can be developed with the intelligence to understand all answers (or supplied data) from any PRODML Server – without requiring any underlying knowledge about the specific server application. In order to achieve this, the PRODML data object schemas must follow defined, understood and agreed-upon rules. The PRODML data schema defines these rules that identify exactly what objects can be returned and their valid characteristics.

Key strategic goals
Key strategic goals of PRODML are to establish a communications and computing infrastructure based on state-of-the-art technologies and standards, with adaptability and flexibility to grow and expand; enable low-risk and low-cost use of proven optimization solutions including interchangeability; and accelerate and encourage innovation in the design, configuration and deployment of optimization solutions.

Application types
The major types of applications involved in optimization programs for which PRODML is being developed are:
•    Operational Modeling. Predictive algorithms for long-term future performance from historical measurements and operational plans;
•    Allocation and Reconciliation. Derivation of critical values from measurements, as in back-allocation of volumes to well bores;
•    Surveillance and Monitoring. Continuous comparison of actual to predicted measurements, out of range condition invokes analysis;
•    Simulation and Optimization. Predictive algorithms for near-term future performance and related set points that achieve pre-defined optimization strategies; and
•    Advisory and Alert. Invoked when post-analytical predicted and actual performance differ significantly; alerts for remedial control changes.

Potential benefits

Improved operational efficiency. Production optimization systems are more reliable and accurate with a lower total cost of ownership because information is used more efficiently and effectively:
Safer operations. Opportunities for remote monitoring, collaboration and timely intervention to solve problems can result in reduced exposure to personnel and a safer working environment.
Improved data trustworthiness and compliance. Improved quality facilitates the management of information to monitor, optimize and report asset performance. Operators can more easily exchange data and collaborate with partners, host governments and service providers, ensuring adherence to contractual, corporate and regulatory obligations.
Accelerated adoption of recommended production optimization practices. Standard processes, as well as innovative variations, can be applied more consistently and efficiently.
Broader awareness of opportunities for production optimization. With access to more timely information, operators can better support field systems and contribute to integrated operations and production optimization.
PRODML work so far has been guided by a series of well-defined use cases. These include:
•    Gas lift well optimization;
•    Free flowing well production optimization using real-time measurements and network models;
•    Field-wide optimization, based on real-time measurements, network models and production forecasts;
•    Downhole measurements, including the traditional flow, pressure and temperature measurements plus other measurements such as Distributed Temperature Surveys (DTS);
•    Integrated use by all application
of an authoritative flow network model;
•    Allocated volumes for smart wells;
•    Production volume reporting; and
•    Injection water handling.

To date, 11 pilot teams have been organized to implement and test draft PRODML specifications based on variations of these use cases, four pilots in 2006 and seven in 2007. Each pilot team carried out a scenario defined by an energy company and used field data provided by the energy company and optimization and support applications and test utility software provided by the software solution providers. A computing “sandbox” with a shared infrastructure and solution provider-specific virtual machines was used to facilitate application interaction via Web Services. Results of the initial four pilot tests were demonstrated at the Society of Petroleum Engineers (SPE) Conference in San Antonio in September 2006,
and subsequent work will be shown at the SPE Conference in Anaheim in November 2007.

The PRODML community is working to define a multiyear roadmap, including use cases of increasing scope and complexity. Among the use cases under consideration are:
•    Longer-term asset planning, including reservoir performance management;
•    Support for ESP and mechanically pumped fields; and
•    Overall well management and optimization.

Decisions about production optimization should be made at the “right time,” whether using real-time, near-real time or historical information. Given the current business climate, the willingness of the industry to collaborate and all of its potential benefits, it seems that now is just the right time for PRODML.