Over the years, the use and application of Neural Networks (NN) has found a “home” in the domain of industrial process control. It is also well known that NN is practically a core function in most popular data mining solutions. 
It is interesting though to note that NN algorithms have been embedded in process control solutions, yet sometimes seen or even projected as a bit of a “black box” or “magic box”. Obviously, because of the complexity involved for most process control engineers to rationalize the output of an NN algorithm, except the performance of the controller.
Root Cause Analysis (RCA) has traditionally been conducted by core statistical applications. RCA is classified based on the use or objectives as:

1. Safety-based RCA, which descends from the fields of accident analysis and occupational safety and health
2. Production-based RCA, which has its origins in the field of quality control for industrial manufacturing.
3. Process-based RCA, which is an “add-on” to production-based RCA, but with a scope that has been expanded to include business processes.
4. Failure-based RCA is rooted in the practice of failure analysis as employed in engineering and maintenance.
5. Systems-based RCA emerged as an amalgamation of the preceding uses, along with ideas taken from fields such as change management, risk management, and systems analysis.

In the course of an RCA initiative, the need to deal with substantial volumes of process data is well known. The combined dependent and independent variables can be in the range of hundreds for a single knowledge discovery problem addressed with data mining, and it is not uncommon that analyses without deep process knowledge can fail because of insufficient understanding of the process characteristics and behavior in question.  
As such, data collection and the ability of a data mining solution to interface, sometimes in near real time, with plant data bases residing in a control systems (i.e. Integrated Control and Safety System) or a Historian data base, is key to the development of an integrated solution that can be deployed for use in a dynamic environment versus performing data mining analytics offline. 
The situation in the Process Industries is that it deals with large data sets, considering that the time resolutions of such sequences can be seconds or even milliseconds.
Take the scenario of a huge data set with a long sequence of events and alarms, with thousands of triggered flags or events, logged operator actions and also changes in battery limit conditions, and then add changes that are being tracked due to heat exchange fouling or catalyst characteristics, rotary equipment, etc. , and then try to rationalize the outcome and impact on the behavior of continuous or discreet variable or a number of variables, e.g. trip events or an alarmed deviation of a safety or quality variable. Try also to visualize the endless number of dimensions (variables) involved as those related to the plant or process areas, the time dimension itself, the operators involved, the process unit operations or physical assets associated with the plant areas and units, etc., and then it becomes obvious that a data mining scenario of high complexity evolves. Yet , this is where data mining brings value and makes its money, because data can be both explored and analyzed in so many ways, but also used for predictive purposes by using the same techniques as in the retail example. It is by far a more complex situation that any other industry can offer.
In the case of the RCA, data, once extracted, transformed and loaded for mining, rules about associations or the sequences of items as they occur in a transactional database can be established and make them useful not only for addressing the RCA problem in concern, but for many other applications, including exploratory and predictive data mining, as for example predicting runaway conditions for a catalytic reactor in a plant or preventing off spec production, or even avoiding trip of critical equipment in a plant. 

The easy part of the answer to this question is: When you can answer difficult questions quickly and accurately. Before answering the more difficult question (how), a bit more about the terminology again. 
It has been a bit of search to come up with a term that represents what Plant Intelligence or Process Intelligence is all about. 
As noted in the July issue of ziconNEWS, Process Intelligence is used for Business Process Intelligence (or BPI). Plant Intelligence may be more of an applicable term, but this term may be too limiting, since the concept is to combine business, process and control system intelligence for related data. 
Some use the term Manufacturing Intelligence, which is more common in the discreet manufacturing sector and not too common in the Process Industries, which prefer to distinguish from discreet manufacturing. 
So, here is a suggestion. How about Process Plant Intelligence. This could cover intelligence derived from everything in any process plant field and control room, as for example Levels 0, 1, 2 systems, to the executive management financial, planning, etc. or as known Level 4 systems.
So, to answer the question fully ... one can also claim that Process Plant Intelligence is in place when you have a reasonable infrastructure in place with Level 1 - 4 type systems. These include a DCS or an ICSS (Integrated Control & Safety System) with a Historian. At Level 4, there must be an ERP system to manage business requirement and a planning system to develop production plans. There also must be a Maintenance Management System (typically part of the ERP). 
At Level 3, there should also be a Data Reconciliation System (DRS) in place for reconciled heat and material balances and possibly a LIMS system. The obvious question is do you need to have an MES? The answer is no, but it will be great of there is one, since more useful operations action - oriented data can be generated and made available. 
Now, with such systems, the plant generates lots of data at all levels. Data which can fuel knowledge. This data may be provided at different levels in the form of various reports. This gives Exploratory Intelligence, but not necessarily Predictive Intelligence. Even for Exploratory Intelligence the "drilll-down" "slice and dice" capability of the data exploitation may be limited, unless properly data management techniques are properly implemented (e.g. OLAP reports). A full Process Plant Intelligence system then that includes data mining can provide answers to difficult questions and predictive replies like:

1. During which operations shift and at which time period a certain Heater was run at its lowest possible cost and highest efficiency?
2. What is the expected load and cost of utilities over the next production time horizon.  
3. During which operations shift we had the maximum deviation between planned production and actual production?
4. What is the likelihood for a trip for a specific group of rotary process equipment.
5. What is the expected run for a furnace before decoking?

  These are some rather questions that can not be answered unless a Process Plant Intelligence initiative has been successfully deployed. The rest of the questions and answers can be endless. And endless are the possibilities that Process Plant Intelligence can offer to those who endorse it for their plants and companies and begin to look more into this domain.