A Object-Oriented Implementation of a Chemical Waste Consolidation Expert System

Return to front page

1. Background and Rationale


2. Interface Design

• 2.1. The Chemical and Chemical Mixture Interface
• 2.2. The Bottle or Transporter Interface
• 2.3. The Drum Storage Interface

3. System Design

• 3.1. System Environment
• 3.2. The Waste Consolidation System Design Overview
• 3.3. The Chemical Library: Class ChemBook
• 3.4. The Bottle (Transporter) Library: Class BottleBook
• 3.5. The Drum Library: Class DrumBook

4. Evaluation and Results

• 4.1. Real-Time Application of Prototype
• 4.2. Software and Logic Testing
• 4.3. Usability Testing
  

5. Future Work

6. Conclusions

Bibliography and References

Appendix A. Example Drum Report

Appendix B. Chemical Compatibility Testing Data

6. Conclusions

The software performs wholesale chemical compatibility testing in real time. Large commercial packages that use molecular modeling and reactivity analysis can perform the task on chemicals for which they have only a structure, but the analysis cannot be done in real time. This feat alone makes the package unique. The costs of chemical waste disposal are rising and expected to rise faster in the future. This package will save the University several thousand dollars per year for the foreseeable future. The chance that the next person responsible for the chemical waste is an expert on consolidation is quite remote. The task is daunting when very few have the total knowledge to deduce how a chemical compound will react in an extremely complicated mixture based on the structural groups on the molecule. While the major proportion of chemical waste has spent its reaction potential and will generally perform in a degraded capacity, the infrequent exception to this general rule can prove disastrous to the facility as well as the consolidator’s health and confidence. The greatest challenge for the current consolidator was having no technical support of which to ask critical questions in real time. The software package can and should fill that void for the next consolidator and will perhaps alleviate some of the anxiety endemic to this task. 

Information generated by this software will be used to insure the University is in compliance with the substantial responsibilities of generating chemical waste. Active, real-time dissemination of the information to the emergency system at the University will increase the safety for many members of the community. It also shows a proactive stance to deal with the inherent danger of these operations. The University is home to the National Spill Control School, whose students attend from a great many of the chemical industries. The efficient, thorough and accurate handling of the chemical waste by the University demonstrates responsibility to these students. They, in turn, can convey this perception and knowledge back to their respective industries.

From a programming standpoint, the software fulfills the requirement for a creative and substantial project. The solution of three abstract classes that represent the physical entities is elegant, intuitive and shows sophisticated knowledge of the object-oriented programming language. The platform performs in the designed task marvelously. Division of the information into objects and structures illustrate that the data structures should mimic the organic items they represent. This was not either apparent or intuitive at the beginning of the software coding. Above all, it is an extremely scalable response to the technical problems of this task. This foundation can be used for the more complex calculations described in Section 5 and can make the software much more powerful. The classes are driven by a core program that provides modular separation to subdivide an extremely complex task into manageable units. Since the software is authored in C++, it is universally portable to any platform available and can easily be shared and cannibalized by other programmers. The experience allowed implementation of the course work theories on software design cycles, and the experience proved that the time frames of cyclic software revision are grueling and unmerciful. Translation of linked lists to perform as a database demonstrates the similarities between the rationale for multiple tables in a database and cross-linked lists in a derived class. Taken as a whole, the project development experience proved surprising in many ways.