Monthly Archives: September 2019

Artificial intelligence development is changing how industry works

Artificial intelligence development is changing how industry works

Category : Blogs

Artificial intelligence development is changing how the industry works

Many industries are going to benefit from artificial intelligence development. It’s hard to say which ones in the long term will find the highest level of success, but we can already see significant benefits in a host of industries.

At its core, artificial intelligence is a tool that can acquire, organize and analyze vast amounts of data to create and parameterize models to recognize patterns and make predictions. AI is delivering many benefits and its continued use is the key to making a business more competitive. By automating some of the repetitive, basic tasks, a company can increase productivity, reduce mistakes and enable quicker, better decisions. In insurance, for example, companies are using AI to automate claims processing. The entertainment industry uses AI to optimize streaming services and suggest content based on an individual’s previous choices and comparing it to the choices of others.

Leveraging AI Today

If you’re a business or a company wondering about what to do about AI, whether to use it or even when to use it, then the answer is, Yes. Businesses must think about using AI. Artificial intelligence is a practical tool, and just like banks use it to prevent fraud or healthcare uses its algorithms to scan X rays, companies should look to solve problems and challenges with AI.

In engineering and manufacturing, artificial intelligence is already enhancing scheduling in a factory by improving downtime and conducting predictive maintenance scheduling. Artificial intelligence saves companies money by reducing costs, for example by collecting data from running machines in the factory and feed it into training for predictive maintenance AI models.

Manufacturers can use these models to detect signs that maintenance is needed, such as changes in vibration signals which might indicate there is a developing problem. They can then schedule a maintenance session at the downtime of their choosing, perhaps overnight on a Saturday where there could be minimal or no loss of production. Naturally, it’s more economical to perform maintenance at the company’s discretion than having an expensive machine offline for several days, while possibly waiting for delivery of replacement parts from somewhere on the other side of the world.

Companies like Electricity Generators that use wind turbines are using AI-trained models to do maintenance and optimize their operations. Without this knowledge, they would continue guessing about how to evolve their operations to greater efficiency. They would also maintain the turbines based on a schedule and not whether the turbine required it, or ad hoc when issues arose. With the help of the data from sensors mounted on those turbines, they can better predict maintenance needs and how to minimize downtime. This makes a significant financial impact and can be the difference between being competitive and not being competitive.

Additive manufacturing, or 3D printing, and simulation are other areas in manufacturing where AI will play a more significant role. Manufacturers will need to simulate and validate to check that an additive manufacturing process will produce a quality part and will integrate within a complex product. As this industry grows, there will be more and more manufacturing done via additive manufacturing. It will open horizons for components and systems that are just not possible to traditional manufacturing techniques. Additive manufacturing needs AI to help improve the design, its validation and the planning of the manufacturing components.

The pioneers in 3D printing are the companies producing software and the components to go into the additive manufacturing machines. This includes using artificial intelligence to improve the digital twins to help manufacturers of everything from cars and planes to the factory or smaller components to make better predictions about their performance. AI can simplify the set-up of simulations, speed up those simulations and detect problems earlier without the need to conduct detailed analyses. AI truly plays a part in all these types of analyses and can help companies use the digital twin more effectively.

Today we’re seeing the finance, advertising and consumer industries benefiting the most from AI, but that’s likely because they have access to data. With lots of information, the leaders in these industries are feeding continuous data into their analytics algorithms, improving their models, and getting deeper insights than before.

As more data becomes available, especially in data-sensitive industries where proprietary technology is subject to privacy and confidentiality, companies will struggle to divulge their data. Healthcare and manufacturing industries are particularly concerned about the confidentiality of their data as it’s part of their competitive edge. For those offering services involving AI, it is imperative to address these concerns and information regarding how it will be used and how protection will be provided.

Three industries where AI is already succeeding

There are countless examples as to the importance of AI and its abilities to help businesses maintain their competitiveness and increase both growth and profits. AI is already being used for data collection, assisting companies to become more efficient and productive.

AI in banking

AI is already being used in the financial industry to help companies reduce fraud, assess risk, improve trade and minimize the costs of making credit decisions. But AI in banking is moving a little deeper. Deutsche Bank, for example, already uses this technology to monitor customer phone calls where they can pick up potential fraudulent cases and alert the operators. It is also detecting possible fraudulent transactions before they happen. This reduces financial loss from fraud, which is often publicly embarrassing for a brand.

AI in healthcare

Artificial intelligence is being used with great benefit in the healthcare industry as well. In the United Kingdom, there’s a serious shortage of radiologists. What AI in healthcare is accomplishing is the automation of analyzing the scans, helping to determine symptoms and even suggesting diagnoses. This both speeds up the diagnostic processes and allows radiologists to better use their time and make smarter decisions by automating preliminary scanning and analysis to eliminate easily detected negative results and focus their attention on the cases where decisions require their expertise.

At Stanford University’s Center for Artificial Intelligence in Medicine & Imaging, researchers continue developing artificial intelligence technology to scan mammograms and reduce false negatives and false positives. Ross Shachter, associate professor of management science and engineering at Stanford University, does not see radiologists being replaced but believes AI can assist in determining breast cancer cases more effectively and efficiently while ensuring professionals can better manage their time with patients.

Artificial intelligence developments are also being used to build medical devices that aid in patient health. Implementing a predictive algorithm that gathers data during the development and production of a medical device requires data and performance records to determine the probability of that device working to the standards necessary for government approval and healthcare acceptance. This gives engineers more time to develop medical devices and manufacturers greater confidence in product introduction.

Artificial intelligence in insurance

Insurance companies have access to an insurmountable level of data – a key ingredient to the successful implementation of artificial intelligence. Using this information makes processing claims, underwriting and even customer service more efficient. If a customer wants to take out a policy or make a claim, they’ll likely interact with a form of artificial intelligence like a chatbot that uses algorithms to address customer concerns. Insurance companies can free up their employees to work on the more complex cases requiring a higher level of human judgment, whereas AI deals with straightforward cases.

With the adoption of machine learning and artificial intelligence, along with the data from the Internet of Things environments, manufacturers and engineers across a variety of industries can increase productivity and reduce operational risk. Predictive and prescriptive maintenance and optimized asset performance management are just some of the ways every industry from healthcare and entertainment, to banking and manufacturing, are utilizing this innovative solution.

Click here to learn more about the future of AI and its uses.


Integrated Quality Tools for Supplier Quality Management

Integrated Quality Tools for Supplier Quality Management

Category : Blogs

Integrated Quality Tools for Supplier Quality Management

Increasing global competition over the past decade has forced automotive OEMs and their suppliers to improve quality and efficiency. There is a growing need for tools that make it possible to measure the quality processes used by each member of the supply chain since the OEMs are committed to providing quality products for their customers. Quality issues arising anywhere in the supply chain may put the OEM’s prestige at risk. Because of this, having quality guarantees with regard to the materials and components provided by suppliers can determine the success or failure of a product.

A structured approach to quality, based on an effective Quality Management System (QMS), is of crucial importance for any organization committed to achieving product and customer satisfaction excellence. Supply chain and quality management are two significant aspects of ensuring a company’s leadership in a dynamic and globally competitive environment. QMS capabilities are particularly important in the automotive supplier network, where requirements should be defined and then transferred through the supply chain. In recent years, the relationship between an OEM and its suppliers has changed. To achieve competitive advantages, the main strategy is to change from a sometimes adversarial relationship to a fully cooperative collaboration between customers and suppliers.

Moreover, suppliers need to be third-party certified to ISO9001 and IATF 16949. The scope of this requirement affects subassembly, sequencing, sorting, rework and calibration services in addition to direct material suppliers. Potential suppliers are released internally after a quality process audit and/or a technical process audit depending on the perceived risks. A proactive approach to Supplier Quality is demanding and decisive in all product life cycle stages, such as:

  • Development Phase. To release and validate the supplier and the bought-out part for the production phase.
  • Production Phase. To inspect the supplied parts by incoming goods control.
  • Production Check Phase. To assess the suppliers and their parts and provide a ranking.
  • Utilization Phase. To track the problem-solving process of the supplier (short-term) and to audit the results of it (long-term).

How does QMS Professional support supplier management?

QMS Professional provides automotive OEMs and tier suppliers with the integrated quality management tools needed to ensure the quality of materials and components provided by suppliers.

Two key modules of QMS Professional enable manufacturers to manage the quality of materials and components from their suppliers. First, QMS Supplier Assessment Management (SAM) helps manage your supplier base for optimal product quality control. With QMS SAM, you will have access to real-time quality data relating to all your suppliers and their incoming goods. A strong and viable supplier network can also help you minimize risks and associated costs while bringing a better quality product to the market.

Second, QMS Professional includes a Supplier Portal within its Concern and Complaint Management (CCM) software. This capability enables you to quickly manage product defect complaints in situations in which the supplier is considered responsible. The designated supplier can access and handle complaints using the eight disciplines (8D) problem-solving methodology. Throughout the process, the customer has access to the current handling status, including changes and updates. With the implementation of the QMS Supplier Portal, you can replace time-consuming manual communication with a controlled, automated system. This automated approach allows you to save valuable time by providing immediate notification of handling status to your customer.


ERP - Making the Investment Work using APS

ERP – Making the Investment Work using APS

Category : Blogs

ERP – Making the Investment Work using APS

The materials war has been won (via MRP and/or Kanban control), the capacity war has just started. Companies with a high cost of raw materials and/or high value of end products are trying to move towards make to order and away from making to stock but in doing this many are finding that a knowledge of the capacity constraints of their plant is key to answering the question ‘when can we deliver?’

What these companies need is a tool that will show them visually what the current load is, what the impact of unexpected events will be, and to be able to ‘what if’ solutions to deliver performance across all orders. They need the equivalent of a crystal ball for their manufacturing operations.

Of course, there are different types of production and ways to supply the demand and for each case, there is a key process and IT considerations for moving to better, leaner and more efficient operations.

For example, companies that are looking to techniques such as Kanban to control the flow of products and materials through a manufacturing process and in many companies, this is very effective for some or all of their plants. The requirement for IT is thus much lower but for many companies perhaps 80% of the parts can be handled in this way while the other 20% cannot so there is often a requirement to handle control differently depending on the part.

However, in an MTO, engineer-to-order, and mass customization company with many products and common resources that make them, scheduling becomes more important. We have seen cases where significant variations in cycle times from one product to another for each stage in a classic pull flow of an assembly line has made it difficult to predict the makespan for a particular batch. This was dependent on what other products shared the same line, their batch sizes and their sequence.

Companies are also increasingly turning to out-sourcing of all the components and just assemble to order. They may have a long order book stretching months ahead but still find it difficult to predict the consequences of a sudden change because they do not have the capacity information of their suppliers to provide accurate answers.

Outsourcing is fine but it can lead to an erosion of a company’s visibility and control of its operations.

Measurement of the benefits achieved from a software implementation is often difficult because the objectives were not explicitly spelled out at the beginning of the project.

Traditionally companies would seek to control their operations better and along the way increase their stock turns and reduce work in process which leads to shorter delivery times. Certainly, delivery performance is becoming more important but sometimes it is difficult to put a quantitative value against it. How many extra orders did we get because of better on-time performance?

Accountants don’t really help either because they will focus on resource utilization and ‘overhead recovery’ rather than on delivery performance. The old saying of Accountants ‘knowing the price of everything but the value of nothing’ comes to mind.

Turning data into information and that into knowledge is a powerful argument and if the ERP company can’t provide what users want then a ‘best of breed’ product must be a good way forward for many companies.

It’s just another way of giving Visibility in what tends to be a ‘black box’ of data that is difficult to get at.

Today we know that most ERP systems have very limited functionality in detailed operational scheduling and APS tools have a proven track record in providing companies with an operational crystal ball to see what is happening and what will happen based on the constraints of their factory.

Knowing what the problem is can be halfway to solving it using the ‘what-if’ analysis available to APS users and knowing what will happen in advance gives the planner a considerable advantage in making timely and cost-effective decisions.

The benefits are again well documented in case studies. These include lower raw material stocks and work-in-process while obtaining increased operational efficiency and on-time delivery performance.

APS brings with it additional tools that allow, for example, both material and resource constraints to be considered and some of these tools include a memory-resident BOM exploder for carrying out ad-hoc capable of promise inquiries in seconds. For many companies, this is the ‘Holy Grail’ that they want to add to their existing ERP installation.

On the other hand, collecting data but then doing nothing with it is just giving you an insight into what happened and maybe why. But using the information to help close the loop between what has happened and what will happen is a more powerful argument for its use.

Right now, in many companies we have ‘islands of control’ whereby a cell, a department even a factory within a supply chain seeks to meet specific objectives that have been given to them and even more important, they are measured against those objectives as key performance indicators, KPIs.

However, what may be a KPI for that unit, cell or department may not actually be in the best interest of the entire business. Companies need to take a holistic view of their business and see what is best and apply those to the KPIs they provide the individual units.

By the Author – Marco Antonio Baptista is the MOM Channel Leader in Americas at Siemens Digital Industries Software.


Crossing the Chasm from ERP to APS

Crossing the Chasm from ERP to APS

Category : Blogs

Crossing the Chasm from ERP to APS

Today we well know that, with few exceptions, ERP systems assume adequate resources are available when required, i.e. resources have infinite capacity. ERP systems have a BOM exploder and inventory control data that typically take orders for products breaks them down into component parts and calculates when to start making them based on the individual lead times perhaps adding adjustments for queuing time etc. No account is taken of the real capacity of resources and whether the resources are overloaded or not the same lead-time is used to calculate the launch time.

Because of that, many production planners and/or schedulers develop their own solution based on spreadsheets to balance capacity and demand. Solutions based on spreadsheets help a lot, no doubt, but it is time-consuming, susceptible to errors and rarely captures the tribal knowledge required to face daily challenges related to the need for re-planning and re-scheduling production due to demand changes, breakdowns, quality issues, material issues, absence of employees, etc.

Available for many years now, APS (Advanced Planning & Scheduling) systems are the solution adopted by companies to enhance ERP functionalities and at the same time, overcome the limits imposed by spreadsheets.

ERP suppliers are increasingly offering finite capacity capability to schedule works orders so that operations are only planned when resources are available. Consequently, materials can be ordered to arrive just in time-based on when they are needed for the operation to be carried out. In this way, it has been shown that inventory levels fall and bottleneck resources are not overloaded. Work in progress is minimized, lead times are more predictable and delivery dates more reliable.

To become successful, three important factors should be observed:

  • Integration of the scheduling system with other applications.
  • Ability to accurately model a plant’s operations (using finite capacity scheduling).
  • Frequent generation of new schedules.

Finite capacity is an integral part of any APS and thus for it to be successful then it must be integrated with other applications like ERP, must have powerful modeling capabilities and run fast enough to re-generate a schedule on a regular basis.

There are two types of APS products on the market. Your choice will have a significant impact on the money you pay and the time it takes you to implement. Some APS solutions subsume and partially replace functionality in your existing ERP system while other solutions enable your existing ERP system to acquire APS functionality.

With the first option, the ERP system has very much a subservient role to the APS with BOM structures, inventory control, aggregation rules, etc. being held within the APS as well as the finite capacity functionality. In effect, similar data is held in the APS as in the ERP system. The latter is left to deal with accountancy matters such as sending purchase orders, invoicing and accounts.

The problem is that data synchronization problems between the APS and ERP packages lead to very expensive and elongated implementation times and in addition, the APS must have client-server functionality since it is replacing the client-server ERP system. In effect, your company is paying hundreds of thousands for replacing what you already have.

Crossing the chasm from ERP to APS using this approach is often beyond the financial and technical capacity of your company to do properly and therefore the system ultimately ends up being abandoned.

The alternative is to use a solution that focuses on the real problem, scheduling, which can use the data supplied to it by the ERP to provide the answer to synchronizing materials supply with a proper finite model of the production environment. This alternative provides the bridge between ERP and APS.

It ‘enables’ rather than replaces ERP and your costs are significantly less to apply and install.

The ERP system retains the BOM structures, explosion, routings and inventory control while the APS deals with the problems associated with the allocation of materials between aggregated orders and uses this as part of the scheduling function so that the constraints include not only the machines, labor, and tooling but also the constraints associated with the availability of materials at raw, intermediate and finished stock levels.

The advantage of this approach is that you take full advantage of the ‘specialist’ knowledge and development focus of the point application providers, the APS vendor, but still use your existing ERP system.

The key to a successful application, as we have already learned, is to have an accurate model of the facility enabling accurate and achievable schedules that can be generated in a reasonable time and fully integrated with other systems. Synchronization of data is much less of a problem since all data comes from one source and communication between the APS and ERP can be done using the de-facto off-the-shelf tools such as ODBC, COM, DCOM, etc. that are available today.

So, when you are considering the ERP to APS chasm ask yourself the simple question:
Is it necessary for me to throw away my existing database and start again or can I find the right application that will fill the hole in the functionality I need?

Filling that hole may prevent you from falling down a much larger one when attempting to cross the ERP to APS chasm.

By the Author
Marco Antonio Baptista is the MOM Channel Leader in Americas at Siemens Digital Industries Software.