Category Archives: Manufacturing

Some Things Should Be Autonomous, But Automobiles?

As per this recent K@W article, automakers are working hard to advance the state of the automobiles, and that’s a good thing, but should the ultimate goal be autonomous, driver-less vehicles?

While park-assist, remote control parking, trailer assist, construction site assistant, blind spot monitor and the pre-crash occupant protection system and other safety systems are a good thing, there’s a difference between adding technology to alert a driver to danger and creating a car that drives itself in an effort to remove the human element. While human error is the leading cause of accidents, it’s not the only cause of accidents. Sometimes an animal, or human, jumps in front of the vehicle or an object falls in front of the vehicle, sometimes something breaks and the vehicle can’t be stopped, and sometimes a natural disaster happens.

If there’s not enough time to stop, a computer is not going to be able to stop the car; if something breaks, a program can’t fix it; and if an unexpected event occurs, will the algorithm know how to deal with it? For example, even if there’s no time to stop, a human might be able to take evasive action and avoid hitting a person who steps in front of the vehicle without warning. But if the only choice is hitting a person or hitting a building, will the algorithm make the right choice? (Cars and buildings can be fixed, dead people can’t). Or will it keep calculating to infinity in hopes of finding a collision-free path of action, and hit the human in the process. What if the failure causes a disconnect between the core processor and the brakes? What if an earthquake happens? Will the algorithm be able to interpret the readings correctly?

But more importantly, what if the system crashes? The average car already has more lines of code in its operating system than in an average computer operating system. As per this article over on the MIT technology review, many cars have a hundred million lines of code in their operating system. For comparative purposes, Windows 7 has about 40 Million lines of code. How many lines of code is it going to take to create an operating system that can drive an autonomous vehicle that performs well enough for a government to consider allowing it on the road? Hundreds of Million, if not a Billion. That’s a lot o code. How do you adequately test that much code? You don’t. You can never guarantee that the code is error free and that system won’t crash. You can only test until the probability is high enough for you to accept as likely to be error free in practice.

And what happens if an unexpected event happens at 70 MPH on the highway and the system crashes? Nothing good.

But the real concern is what happens when the OS is hacked? Your computer gets hacked, you lose your personal and confidential information and the computer becomes inaccessible to you until you unplug (and reboot) it. If your car gets hacked, it becomes inaccessible to you until you cut the power. You can’t do that at 70 MPH, and since all cars are being built with 4G, bluetooth, wifi, etc. — if a hacker gets control of it while you are on the road, he can crash your car into another and there will be nothing you can do.

And if the hack is the result of a bug in the OS that allows for a massive zero-day exploit, a hacker could take control of all cars on the road on the same communication network, and cause them all to accelerate until they hit something. If tens of thousands of vehicles were hacked and subverted all at once in a zero-day exploit, the widespread damage that could be caused would be hundreds or thousands of times worse than most terrorist groups currently achieve when they manage to hijack a plane or blow up a single building.

In other words, removing the human completely from the picture doesn’t increase safety, it decreases it. If we must have autonomous vehicles, then they better all come with an old-school code-free manual override switch that, in an emergency, will let us turn the computer off so we can drive home safely and tell those darn kids to get off our lawns.

Machine-to-Machine Networking Can Take Predictive Analytics a Long Way

… but the day the machines can figure out that sunlight through a window is causing the machine to malfunction is the day the machines take over and kill us all!*

What brought on this rant? A recent piece over on ThomasNet on “Machine-to-Machine Networking” that said thanks to predictive analytics, BMW found that the bug wasn’t in the machinery; sunlight coming through a window in one facility was slowly heating up machinery to temperatures beyond optimum range and affecting the components being produced. While this is correct, it is misleading. The reality is that thanks to embedded sensors and M2M networking, the analytic software that powered the predictive analytic engine noticed that the core temperature in part of the machinery producing cylinders on one production line increased during the day while the core temperature of the same machinery on an identical production line (in another plant) stayed constant once the machine heated up. Since it’s generally a bad thing when a machine overheats, the software alerted the production manager that the machine was running too hot late in the day and probably needed to be serviced.

At this point, the production manager would assign an engineer to inspect the machine and run some basic tests, only to find that everything was working fine. However, armed with the data that the machine was overheating, upon finding no probable internal causes, the engineer would examine the surroundings, particularly at the time when the machine typically overheated, and notice that it was in the direct path of sunlight later in the day. Since engineers know that light is a heat source and that metal observes heat, especially when it is in the path of direct sun for hours, the engineer would conclude that at least part of the problem was the direct sunlight, shield the machine, and monitor the performance, and core temperature, for the next few days. At this point, the engineer would notice that the core temperature and production quality stayed constant and would then be able to conclude that the sunlight was the cause of the problem.

All the M2M-enabled predictive analytics package for preventive maintenance is able to determine is that something is not operating at typical performance levels, be it heat, throughput, energy usage, etc. It points you to the source of the problem, not the root cause. You’ll still need a smart engineer to figure out why the machine is overheating, why the energy usage has shot up, why the defect rate is increasing, etc. In a few cases, the software will be able to determine that a sensor is broken, a connection is down, or a part is broken when data cannot be retrieved, checksums are incorrect, or scans come back with known error types. But this is not the typical behaviour. On average, the best you’ll be able to figure out is that a part needs to be replaced, but not what’s wrong with it. In many cases, it will be cheaper to just swap out the part then to try and diagnose and fix it, so you’ll do just that and not worry about what went wrong. It’s a valid approach, as it keeps the machine up, costs down, and saves you money in the long run — but not one that helps you figure out why the part wore out. Was it a defect in the part or a problem with the machinery it’s embedded in? If the former, you don’t care what the defect was — only that the supplier replace it under warranty. If it’s a problem in your machinery (such as voltage spikes causing the part to burn out early), then you’ll (eventually) suffer multiple part failures and want to know the root cause (but you won’t be able to even suspect the machinery until you have the 3rd such failure).

the doctor is all for predictive maintenance and using sensors and machine-to-machine networking to be efficient and cost-effective about it, but wants all providers and promoters of such technology to be very clear about what it can, and can’t do. It can detect variances (and abnormal operational conditions) and correlate them to patterns that suggest potential problems and a need for part replacement, but it can’t say for certain why those problems exist. In the hands of a smart engineer, it will help to diagnose the machine or part that is source of a rare or difficult problem much faster than the engineer could track the source machine or part on her own, but it won’t be able to identify the root cause. That will still require brainpower. The machines can’t replace us yet. Remember that before you get oversold.

* Unless the machines need us to power The Matrix. The scary thing is that the day of dread may not be too far off! The NSA is building Skynet, Amazon is building autonomous drones, GE and Boeing are trying to make everything smart, and 3-D printing is at the point where we can now make primitive replicators. And everyone seems to have forgotten about Asimov’s three laws of robotics, thinking AI is still decades off when you can buy a GPU on a high-end PC graphics card that can do over 4 Trillion instructions per second. (In comparison, peak performance of the Intel 8088 processor was a mere 1 Million Instructions Per Second.) The computational power exists — all that is missing is the algorithm.

Apparently Accountants Have a Very Different Meaning for the Word Enormous

According to a recent article in Modern Material Handling (MMH), which reported on the Grant Thornton Realities of Reshoring Survey and quoted Wally Gruenes, Grant Thornton’s National Managing Partner for Industry and Client Experience, the results (of the survey) could dramatically impact U.S. trade balances, and should provide an enormous boost to domestic manufacturers, retailers, wholesaler/distributors and service providers. Great news, right?

Let’s dig in. According to the results of the survey, more than one-third of U.S. businesses are likely to move goods and services back to the United States in the next 12 months. In particular, 42% of executives indicated they were likely to bring back IT services, 37% said they were likely to bring back components/products, 35% said they were likely to bring back customer services or call centres, and 34% said they were likely to bring back (raw) material. Not exactly enormous, but not too shabby either. For one third of companies to at least be thinking in the right direction, that’s pretty good. Except when you dig in and realize that the numbers imply that as much as 5% of overall U.S. procurement may come back to the United States. 5% is not enormous! It’s not even close. And this is the best case scenario, which we know isn’t going to happen.

First of all, someone would have to get off of their @ss and push for a major change (and in your average company, meet a lot of resistance). This is something that only happens in market leaders, which we know are only (depending on which analyst firm you ask) the top 8% to the top 20% of the market. Secondly, a C-Suite executive, still focussed on quarterly numbers and penny pinching, would have to sign off on what could be moderately high one-time expenses associated with re-shoring — expenses which would be minimal in the mid-to-long term, but which would probably really irk the CFO in the short term (and mess up his attempt to look good for Wall Street). (And given the number of companies that have invested in training over the last 5 years, even though case studies from Procurement training institutes, including Next Level Purchasing, have proven ROIs of 10X to 100X from proper training investments, we know that few companies in North America put long term savings ahead of short term gains.) Thirdly, someone has to be willing to get a little egg on their face and admit that maybe outsourcing (so much) to China wasn’t that great of an idea in the first place — that if appropriate investments had been made at, or near, home to increase productivity, decrease production time (and cost), and improve operational sustainability, similar cost savings could have been made over the long term with an appropriate investment up front. How many pompous C-Suite executives in North America are willing to fess up and admit they were wrong? (Let’s put it this way, the Mad Men would be an awful lot poorer if more were.)

Long story short, if even 1% comes back this year, the doctor will join you in the dance of joy because he just doesn’t see it happening. He’d like nothing more than for 10% to come back, especially since he’s been preaching the importance of Home Cost Country Sourcing since 2007, but believes only the true market leaders will take any actions at all. Most companies just aren’t hurting enough to bother.

aPriori, rationi viam ad sumptus! Caput II

In yesterday’s post, we re-introduced you to aPriori, the masters of Enterprise Product Costing that have been working their cost reduction magic for a full decade, taking out mountains of cost before the first part is produced! We noted that, even though it’s been over half a decade, the masters of costing have stayed the course and are still focussed 100% on taking cost out during the design and production phases, where up to 80% of the cost of a product is locked in. They do this through complex process models, built on CAD geometry, that they embed in sophisticated VPEs (Virtual Production Environments) which are populated with accurate cost data for each material, machine, and overhead factor that contributes to the total production cost.

Today we want to highlight the major improvements made in the last five years.

Significantly More Production Process Models!

When SI first reviewed aPriori, their out-of-the-box capabilities were limited to metal-based parts only, and there were only a few dozen process models. Now they can handle virtually any metal and plastics component you can think of and support over two hundred production process models out of the box. In addition, they recently signed some very big name electronics manufacturers and are adding electronics process models to their repertoire, and a few of these will likely be available out-of-the-box this year.

Significantly More Virtual Production Environments!

Now that they have close to 100 customers across the Americas and Europe, that produce their components across the Americas, Europe, Asia, and even Africa, they have up-to-date cost models and accurate VPEs for every major geography out-of-the-box. An engineer, or buyer, can get a rough idea of production cost for any supported production process in any geography before even engaging with a supplier, who can, of course, provide even more accurate cost data specific to their factory.

Support for Every Standard CAD File Format and Just About Every CAD System

The more customers you get, the more CAD systems and file formats you have to work with. At this point in their evolution, aPriori now supports every standard CAD file format and every major CAD system currently in use in the manufacturing sector.

Improved UI

It looks better, responds faster, and integrates the best of CAD and OLAP. The main screen has three sections: the component view, the cost model, and the process model. Each displays the high-level information, but in each the user can drill down as deep as she desires.

Full Excel Export Capability

Not only can the user copy and customize process models and VPEs, update / override any cost, and save any scenario – but they can also export the full scenario and underlying cost model to excel for analysis, review, and distribution.

Powerful Comparison Reports

The user can compare multiple process models, and associated costs, for a part side-by-side, and, if desired, export the full comparison report to Excel.

Roll-Ups and Automatic Process Model Generation and Solution

A user can create a component-based production should-cost model that rolls-up the production should-cost model for each part and the system will automatically cost the full component using the individual part geometries and identified (or default) production processes and, if desired, the lowest cost production process for the entire component.

The improvements save their customers millions every year. For example, the construction equipment manufacturer that saved over 500K annually just on frame and door production also saves over 200K annually on cage rear pivot production. The manufacturer thought that machined casting w/x-Ray was the best way to produce the part, but the aPriori solution was able to determine that a two-step process that first burned the part farm from plate and then machined holed the cavities could reduce the cost from 16.56 to 10.05 on 22K cage pivots per year.

And it’s not just construction equipment manufacturers that save. Thermo King, which produces temperature and climate control products for the transportation industry, analyzed 5.679M in annual spend across 294 sheet metal parts and quickly identified a potential savings of 900K (16%) and realized 400K of this in just 12 days! And a a 6.5B manufacturer of commercial trucks that analyzed 7.7M Euro in spend across 86 sheet metal parts was quickly able to identify that 17 of the 86 parts were “outliers” (and nowhere near expected costs) and through additional analysis was able to identify better production methods that led to a confirmed savings of 1.6M Euro (21%).

It definitely helps to know your expected production costs aPriori!

aPriori, rationi viam ad sumptus! Caput I

When we last covered aPriori in 2007 and 2008 in aPriori and The Sourcing Maniacs 2008 Vendor Tour Part III, they were very focussed on Enterprise Cost Management (ECM) and taking cost out of the design phase. Fast-forward six years later, and nothing has changed, except, of course, the depth, breadth, and usability of their platform — which has grown in leaps and bounds.

Unlike traditional sourcing applications, including advanced spend analysis and decision optimization, that are limited to component cost-based should-cost models, aPriori can also factor in design and production factors to model the full production cycle of the part you are buying (if it’s metal, plastic, or, in some cases, electronics-based) and give you a true understanding of what the part should cost to make. The reality is that the cost of a part is dependent not only on its design, but on the production process employed. As noted in our first post, a supplier that’s always made a certain part a certain way might not realize that new technology or materials would allow them to make that part significantly cheaper if they used a different process. Since the aPriori application instantly and directly interfaces with your CAD program and interrogates the solid model to extract the geometric cost drivers, the aPriori application can automatically determine all the process routings that can be used to make the part, compute the costs associated with each step based upon standard machine, material, and labor costs, and compute the total cost of each part on a per unit basis by factoring non-geometric cost-drivers such as production volumes, the selected supplier or factory set-up selected, and the exact routing and machines used. This is because the aPriori application currently supports over 200 out-of-the-box process models in over 12 major process groups (including, but not limited to, Bar & Tube Fabrication, Casting, Forging, Machining, Plastic Moulding, Powder Metal, Roto & Blow Moulding, Sheet Metal Sheet Plastic, Stock Machining and Rapid Prototyping.

In addition, because the application supports the creation of complete VPEs (Virtual Production Environments) that encapsulate the production processes, a customer can fully model the production and overhead costs associated with each production process supported by a factory in question, including local labour, power, maintenance, and other overhead costs to create a fully accurate should-cost production model, which can be compared to alternate production processes in the factory and other factories modeled with an appropriate VPE. This allows for the true identification of the lowest cost because, as the Sourcing Maniacs documented in their vendor tour post, the COGS is a combination of raw material costs, labor costs, production overhead costs, and margin and these costs not only vary by locale and production process, but in their interaction. For example, just because you identify three ways to make a part and each requires three steps, this doesn’t mean that each process is going to be roughly equal in cost. Not only do different processes require different amounts of manpower or energy (for energy-intensive equipment like lasers, etc.), but reordering the steps can change the manpower or energy required in subsequent steps.

Let’s take, for example, the production of the main Frame sides and door for a piece of heavy machinery construction equipment. An aPriori customer was cutting the entire frame using a laser process. While this seemed efficient, as only one piece of machinery was required, cutting the entire frame and door using a laser cost them 75.54 per frame and door combination, and they required over 14,000 of these combinations a year. That’s over a million dollars on just one part! If, however, as discovered by aPriori who analyzed the geometry and ran it through every possible production process that was available to the manufacturer, they switched to a two-stage production process that involved an initial laser cutting of the frame and door followed by an NC Punch process to punch out the internal cavities, the time required to produce a single frame and door combination decreased by 14 minutes and the cost decreased by 56% to 33.29 (as laser cutting is expensive compared to NC punch).

So what’s new with aPriori? Come back for Part II.