We offer a checklist that will help determine whether your company is ready for digital optimization.
1. Digitization
The term "digital optimization" itself implies that the enterprise must have enough data that can be used to optimize the process. Production data is generated by equipment, various sensors (temperature, pressure, speed, flow rate, etc.), factory laboratories (producing control of products and raw materials). Also, a large amount of data is generated by service units (economists, logistics, energy, etc.). In order for data to be used for optimization, it must be collected and stored somewhere. Ideally, the repository should be one, but this is not necessary: if there are different accounting systems, each with its own database, then you can also work with such data.
2. Stationarity
Digital optimization of unique projects is very difficult: data analysis is designed to find patterns in constantly repeating processes. Therefore, continuous stationary processes are best suited for optimization. Continuously repeating processes are also well optimized. But even inside the production of unique products, there can be a plant or a process that works continuously (for example, in a construction organization there can be concrete production) and such processes (if they are digitized) can also be digitally optimized.
3. Complex physics (chemistry)
Technological processes and installations in which complex physical or chemical processes take place are difficult to conventional modeling. A vivid example is the mixing of petroleum products to obtain a mixture of the required quality: all the parameters of the mixture have a very difficult dependence on the parameters of the starting components. In such places, there is usually some kind of instruction for an approximate calculation and further control and adjustment (when new raw materials arrive, the operator calculates the relative amounts of components according to the instructions, then the mixture is sent for analysis and corrects the quantity based on its results). Such processes involve significant losses of time, raw materials, energy, labor, etc. Machine learning algorithms can create a more accurate model of such a process, which will help the company save a lot of resources.
4. Lots of options
A person does a good job of optimizing a process that has one parameter (for example, you need to select the temperature at which the process proceeds best). If there are two parameters, then it is much more difficult to optimize: the number of possible options is equal to the product of the number of different values of one parameter by the number of values of another. # Nbsp;
In real processes, the number of parameters can be in the tens. The human brain, in principle, is not able to keep such a quantity of data in mind. # Nbsp; Instructions from equipment manufacturers and the many years of work of technologists allow us to maintain the efficiency of the process within an acceptable framework, but in such places, there is always the potential for optimization and digital optimization brings very good results.
5. The human factor
The situations described in the previous two paragraphs lead to the fact that key employees are always present at such complex plants, who not only know the technology but have extensive experience and intuition in managing these technical processes. Transferring experience and intuition to another employee is a non-trivial task, and often simply impossible. It turns out the dependence of the whole plant on one or more key employees.
But the main problem is different. The employee who controls the process by intuition is sure that his decisions are optimal. Meanwhile, this is far from the case. Artificial intelligence always has a potential higher than humans. Therefore, all the places where such key employees are located are very well optimized through data analysis.
6. The initial optimization
Naturally, switching to digital optimization makes sense only where conventional optimization is carried out, i.e. in those industries where attention is paid to efficiency. A good indicator can be implemented one of the quality management systems (6 Sigma, Kaizen, ISO-9000, theory of restrictions, etc.), the use of lean manufacturing; however, even if none of these systems has been formally implemented, but the technical staff is regularly tasked with at least increasing the productivity of the line, it means that the production order is kept to a minimum and data analysis can have an effect.
We hope that after reading the article you have a general idea of whether it makes sense to harness digital optimization. In any case, contact us, we will help you figure it out, conduct a preliminary free examination and answer all questions.
Who should do digital optimization?
We offer a checklist that will help determine whether your company is ready for digital optimization.
25 February 2017
- Comments
Loading comments...