Artificial intelligence and expert systems

AI (for an exact definition, see Section 9.3.2.1) is in reality a cumulative denomination for a large body of techniques that have two general common traits: they are computer methods and they try to reproduce a nonquantitative human thought process. General AI topics are not addressed here. For information on this topic see the various monographs giving fundamental information, including Charniak and McDemmott (1983), Drescher (1993), Rich (1983), and Widman, Loparo, and Nielsen (1989).

The applications we will deal with in this chapter are related to a smaller subset of general AI techniques: the so-called knowledge-based systems, also called expert systems. Referring the reader to Section 9.3.2.2 for definitions, will say only that an ES is an AI application aimed at the resolution of a specific class of problems. Neither a computer nor an ES can think: the ES is a sort of well-organized and well cross-referenced task list, and the computer is just a work tool. Nevertheless, ES (and in general AI techniques) can result in efficient, reliable, and powerful engineering tools and can help advance qualitative engineering just as much as numerical methods have done for quantitative engineering.

ESs have many benefits. They provide an efficient method for encapsulating and storing knowledge so that it becomes an asset for the ES user. They can make knowledge more widely available and help in overcoming shortages of expertise. Knowledge stored in an ES is not lost when experts are no longer available.

Introduction

The keystone of the design of a thermal system or any other chemical process is the process simulation program. Quoting Cox (1993), “All correlations, theories, and physical property data must be translated into computer terms before they truly become useful for chemical process design applications. Data banks and associated physical property models form the heart of any computer simulator calculation.” The quality of these data modules can have extensive effects. Inaccurate data may lead to costly errors in judgment whether it is to proceed with a new process or modification or not to go ahead. Inadequate or unavailable data may cause a potentially attractive and profitable process to be delayed or rejected because of the difficulty or impossibility of properly simulating it. Even the most sophisticated software will not lead to a cost-effective solution if it is not backed up by an accurate database. The ability to effectively conserve energy in many processes is related directly to the accuracy of the physical and thermodynamic data available.

A number of projects, frequently funded by consortia of companies and government agencies, exist for the collection and evaluation of data concerning particular areas of interest: chemical processing, electric and gas production and distribution, biology, medicine, geology, meteorology, music, demographics, etc. This chapter will address the phase equilibrium, thermodynamic data, and physical properties generally required in the design of thermal and chemical processing systems.

Introduction

In today's industrial practice, many plants are suffering from process designs that were completed in times when energy conservation and pollution prevention were not yet key issues. Such plants were built with little regard for reuse of utilities such as heat or water, because these came at low cost and could easily be discharged. Recently, however, environmental protection policies and the awareness that resources for heat and water are limited have led to considerable efforts in the fields of process design and process integration.

In a wide range of industries, Pinch Analysis is accepted as the method of choice for identifying medium- and long-term energy conservation opportunities. It first attracted industrial interest in the 1970s, against the background of rising oil prices. An engineering approach, but one incorporating scientific rigor, Pinch Analysis quantified the potential for reducing energy consumption through improved heat integration by design of improved heat-exchanger networks. Since then, Pinch Analysis has broadened its coverage to include reaction and separation systems in individual processes as well as entire production complexes. Today's techniques assess quickly the capital and operating-cost implications of design options that reduce intrinsic energy consumption, determine the potential for capacity debottlenecking, improve operation flexibility, and, most recently, quantify the scope for reduced combustion-related emissions. In addition, the same basic concepts have been extended to mass-transfer problems to give similar insights into waste minimization, with significant environmental implications.

In this chapter, the fundamentals and some extensions of Pinch Analysis are briefly reviewed.

The importance of developing thermal systems that effectively use energy resources such as oil, natural gas, and coal is apparent. Effective use is determined with both the First and Second Laws of thermodynamics. Energy entering a thermal system with fuel, electricity, flowing streams of matter, and so on is accounted for in the products and by-products. Energy cannot be destroyed – a First Law concept. The idea that something can be destroyed is useful. This idea does not apply to energy, however, but to exergy – a Second Law concept. Moreover, it is exergy and not energy that properly gauges the quality (usefulness) of, say, one kilojoule of electricity generated by a power plant versus one kilojoule of energy in the plant cooling water stream. Electricity clearly has the greater quality and, not incidentally, the greater economic value.

Preliminaries

In developed countries worldwide, the effectiveness of using oil, natural gas and coal has markedly improved over the last two decades. Still, use varies even within nations, and there is room for improvement. Compared with some of its principal international trading partners, for example, U.S. industry as a whole has a higher energy resource consumption per unit of output and generates considerably more waste. These realities pose a challenge for maintaining competitiveness in the global marketplace.

For industries where energy is a major contributor to operating costs, an opportunity exists for improving competitiveness through more effective use of energy resources. This is a well known and largely accepted principle today.