SPAR Modeling Platform - Predicting Asset Performance
Clockwork Solutions develops, markets, customizes, and supports SPAR™ - the most advanced modeling development platform for effective prediction and management of performance and life cycle costs of asset-intensive, complex systems.
Since 1986, SPAR-based applications and solutions have been serving the predictive modeling needs of very large organizations and enterprises worldwide across a number of key industries, including: Defense, Energy, Process and Transportation.
Creating a Realistic Model
When designing a new system, considering major changes in an existing system, or facing demanding operational and logistic requirements, the ability to effectively predict system behavior is crucial.
With the SPAR development platform, users create realistic models of complex systems - quantifying the effects of design changes, different operational scenarios, spare part availability and preventive maintenance schedules on system performance and cost.
SPAR offers a method and tool for describing the fixed and the variable characteristics of a system, together with its operational and logistical environments, in one integrated model. SPAR uses this model to predict the system's performance and cost, addressing variable factors such as:
- The structure of the system, including redundancy and reconfiguration
- The objective of the system, such as throughput, availability, or safety targets
- System operating policies
- System component characteristics, such as initial age, aging rate, and repair time
- External conditions, such as weather or power loss
- Spare part stocking policies, including supplier lead time and part cost
- Preventive maintenance schedules - resources required and their effect on the system availability
Through simulation, SPAR establishes a relationship between these factors and the system's overall performance. A SPAR simulation provides a prediction of a system's performance over time, the cost of achieving that performance, and the sensitivity of system performance to the variable factors affecting it.
Predicting the Realistic Future Behavior of an Industrial System
It is scientifically recognized that a model will never fit reality in all its infinite details, but will rather be at some "distance" from it (known as Distance to Reality, or DTR).
SPAR strives to reduce DTR to a bare minimum, basing models on a rich set of data about the structure of the system, the interactions between the elements of the system, the interaction of the system with its environment, and data related to the properties of the system components and how they move among their possible states. A SPAR model of a capital asset describes the configuration of the asset, statistical properties of each component, contingencies that govern the behavior of the asset, and rules that define the behavior of the support infrastructure.
Once the SPAR model is established, a set of questions is generated relating to the future performance of the system - its failures, spare parts, repair teams, inspections, maintenance, production and anything else that is of interest.
The SPAR model then predicts the future behavior of the system for a given scenario of design, operations, and maintenance parameters. A SPAR analysis evaluates the model under several scenarios to identify the most appropriate configuration, operating policy, and maintenance practices to achieve performance and cost objectives.
SPAR uses time-dependent Monte Carlo techniques - originally developed and applied to model phenomena in nuclear physics - to simulate the behavior of the modeled system. During a simulation, SPAR samples the statistical attributes of the components that make up the system to create an event. It then evaluates the operating contingencies and support infrastructure rules to determine the response to the event. The occurrence of each event and its consequences are recorded in an internal database. This process is repeated several times to create a statistically meaningful database of system behavior, which is later used to predict future performance and system support costs.
SPAR Monte Carlo techniques are optimized for capturing uncertainty during the modeling of very dynamic environments. The Monte Carlo technique provides a basis for assessing the impacts of uncertainty, and therefore delivers a higher level of confidence in predicted performance metrics.
Easily Customized for an Exact Fit
More than just a single piece of software, SPAR is a model development platform, on which Clockwork experts and our clients build custom SPAR-based applications and models.
SPAR technology consists of numerous proprietary and novel extensions to traditional Monte Carlo techniques. These extensions were designed to model real-world phenomena such as age and aging, capacity, demand profiles, time delays, induced behavior, maintenance, inspection, spare parts, repair, and overhaul.
Using the SPAR platform, Clockwork first creates a mathematical computer model of the client’s system, and then develops a custom GUI to help our users best understand and leverage the model. These custom GUIs allow users to enter additional data to expand and refine the model, and then to run additional simulations that account for this new information.
The end result - a predictive modeling solution customized for exact industry-specific end-user needs, which still allows easy on-the-fly changes for optimal performance and results.
For more information:
Also download the following selected documents from our Information Center:
SPAR - Product Introduction, SPAR v5 Product Announcement, The Value of Proper Modeling in Performance and Resources Assessment, Using RMA Monte Carlo Analysis as Part of the Systems Engineering Design Process, The Monte Carlo Method vs. the Analytical Methods
Home |
Company |
Applications and Solutions |
Technology and Products
Industry Solutions |
Information Center |
Contact Us