As the reader of this document may be aware, CARB (the California Air Resource Board) has mandated that all auto manufacturers that want to continue licencing new cars for sale in the state of california shall produce a certain percentage of their units, not less than 2000? vehicles (note: find exact stats and a web link) for use in fiscal year 2004.
This means that a number of automakers will be manufacturing production quanities of electric vehicles.
While most of the technologies involved in a EV are reletively simple and well researched, there are a few areas that have not been given appropriate development. In particular, the embedded computer systems on EVs today are primitave at best. While the technology exists to build a system that will permit cell-by-cell monitoring and balancing of the large agragate cells (12V usually) that make up a EV battery pack, currently no such systems are in place. One company (find name and link) is promising a beta monitoring system in Q2 of 2000, but it will not include software for charge monitoring and failure mode detection, nor will it permit cell balancing. Another system permits cell balancing, but requires substantial financial outlay as each module costs ~$40. This system does no monitoring.
It is essentual to battery pack reliability that a computer be able to see voltage and tempurature of each individual cell. As well as detecting immediate failure modes like overheating battery posts caused by poor connections, such a system could detect long term failures like a single cell increasing resistance or reducing output capacity. Not only could the system notify the end user as to the failure condition, it could also notify the mechanic as to which cell needed service, and what the most likely problem was.
Our first goal is to build such a system - one which can be manufactured for $500 (CPU and associated hardware) plus $15/cell (parts & labor, assuming standerd automated production facilities).
Our second, and more important goal is to write the associated software that goes with such a system - software that can detect cell failure modes and react to them, and software that can control onboard charging hardware so as to keep the battery array in top shape.
Any operater of a series string of batteries is a potential customer - this system, while intended for electric vehicles, could easily also prove useful in a wide variety of other applications, including backup power arrays for UPS units, solar power nighttime storage arrays, and a host of other series-DC-string appliactions. [ insert blurbs about price of oil, increased EV efficiency, etc ] In order to develop clear recognition of cell failure modes, it is advisable to actually observe some cell failures. For that reason, we would like to manufacture a small fleet of electric vehicles (50-100 units) and place them with a variety of customers. They would be sold to the customers at cost, or at a slight loss - the customers would sign a agreement that permitted the company to connect to their vehicles on regular intervals to collect logged data about battery behavior. In exchange for being beta testers of the system, and permitting their battery data to be analyzed, the customers would receive free service on their vehicles for the duration of the test period. [ insert bit explaining that buying a new 'stripped' frame and converting it to EV would cost ~ $22k a unit plus $4k labor ] [ insert bit explaining cost of leasing a production line ] [ explain profit scheme ]