Error "matrix is singular"

Question

I keep receiving the message "Matrix is singular" while attempting the calculation.

I am using a database where I imported agribalyse and all the free available databases (ELCD 3.2, Needs, USDA, Impact World+, OzLCI 2019, Worldsteel 2020, Exiobase, BioEnergieDat).

Since this error can be caused by rows or columns linearly dependent, I examined my process, but i couldn't find the error, so I downloaded the matrix, but i still can't find the 0 values.

I also read that it could be related to the EF, but i don't seem to have used it.

The main problem is that my process is a complex cosmetic formulation, and I had to create the processes to synthetize most of the ingredients starting from the base.

How can I fix this problem? Is there another way?

Thank you,

Bianca

Answer 1

Hi, well for a start, it is maybe not good to import all these databases in one openLCA database that you plan to work with. You can make a validation of the openLCA database to maybe spot import errors first. The 0 rows and columns are one common example that cause a singular matrix, but there are others. When you have exported the matrix, you can even in Excel but also in R calculate the determinant for this matrix (see for excel e.g.https://real-statistics.com/matrices-and-iterative-procedures/determinants-and-simultaneous-linear-equations/), taking in more and more rows / columns, and at one point the determinant will become 0. The last included line is then responsible for the singular matrix. I hope this helps, best wishes, Andreas

Comments

Thank you for your help.
I simplified the database importing only Impact World+ and Agribalyse.
Another solution i found was to attempt the calculation of the impacts for the inputs of my cosmetic formulation until I received the error "matrix is singular"; i discovered that the problem was an ingredient and I solved it.

---

very nice to hear! Thank you for reporting back.